Iowa State University

Digital Repository @ Iowa State University

Retrospective Theses and Dissertations

1949

A piezometer method of measuring soil

permeability and application of permeability data to
a drainage problem
James Nicholas Luthin
Iowa State College

Follow this and additional works at: http://lib.dr.iastate.edu/rtd

Part of the Agriculture Commons, and the Soil Science Commons
Recommended Citation
Luthin, James Nicholas, "A piezometer method of measuring soil permeability and application of permeability data to a drainage
problem " (1949). Retrospective Theses and Dissertations. Paper 14091.

This Dissertation is brought to you for free and open access by Digital Repository @ Iowa State University. It has been accepted for inclusion in
Retrospective Theses and Dissertations by an authorized administrator of Digital Repository @ Iowa State University. For more information, please
contact hinefuku@iastate.edu.

NOTE TO USERS

This reproduction is the best copy available.

UMI

A PIEZOMETER METHOD OP MEASURIIG SOIL

PKRMEABILITY AID APPLIGATIOI OF PERMEABILITY DATA

TO A DRAIIAGfi PROBLEM
by
James Micholas LutMa
A fliesis Submitted to the
Graduate Paeiilty for the Degree of
DOCTOR OF PHILOSOPHY

Ma^or Subject;

Soils

Approved;
Signature was redacted for privacy.

Signature was redacted for privacy.

Head of Major Department

Signature was redacted for privacy.

Dean of 'Graduate Ooliege

Iowa State College
1949

UMI Number: DP13349

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.

UMI
UMI Microform DP13349
Copyright 2005 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company

300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346

S59f
L317P

-ii-

fABLl OF COITEHTS

I.
II.

Page
1

IlfRODUCTIOI
RE?IEf OF LITSRAfURE
A.
B.
0.
D.

Darcy's Law
Equation of Flow
PerBieability Units
Methods of Measurement
1.

Laboratory metliods
a.
b.

2.
3.

Unlined wells
Lined wells
tracers

Factors Affecting Soil Permeability

1.
2.
3.
4.

F.

The disturbed sample

The undisturbed sample

Indirect methods .
Field methods
a.
b.
e.

S.

Entrapped air
Microorganisms
Salts
Temperature

Analytical methods
a.
b.

2.
3.
4.
5.

11
11
11
12
12
12
13
14
15
15
17
20
20

Solution of Flow Problems

1.

4
7
9
10

Bapuit-Forchheimer theory
Kirkhaai-Gardnex approach .

Graphical methods of Dachler, Caaagrande,

and Forehheiaer
The hodograph
Membrane analogue
llectrical analogue
lumerical methods

T "^0 40

21
21
21
25
26
27
28
29
31

-iiiPage
III.

THl PROBLIM
A.

Piezometer Method for Measuring Soil

PeriaeaMlity , . . .
1.
2.
3.
4.
5.

B.

36

Field procedure
Field tests .
Laboratory procedure .....
Results of field tests ..........
Laboratory results

Application of Permeability Data to a Drainage

Problea
1.

b,

36
40
45
50
68
73

Procedure
a.

IV.

36

74

Draimage of soil with uniform

perraeabillty . . . . . . . . . . . . .
Dralmge of soil with non-uniforw
permeability

77
84

GONGLUSIOIS AID SUMMAET

96

LITSIATURI OlfED

97

1.

ACIIOfLED^ffilT

.103

-i-

LIST OF TABLES
Page

xao.

Relation of Preesmre to Flow Through GlaySeelhMai (60).

Field Test Data

51

Field Tftst Data

52

Field Test D.ta

53

Field Test Bata

54

Field Test Data

56

Field Test Data

57

Field Test Data

59

Field Test Data

61

10

Field Test Data

62

11

Field Test Data

63

13

Results of luraerieal Analysis

94

-V-

LIST OF FiaURlS
Fig.

Page

Illustrating Darcy*e Law.

Changes in Peraieability of Soils During Long

Submergence. Allison (l),

18

PermeaMlity-fime Curves for Hanford Loa

Under Prolonged Submergence. Allison (l).

18

Dupuit-Forchlieimer Theory of Soil Drainage.

24

Points on let

34

Relaxation Pattern.

34

Piezometer Method of Measurement of Permeability.

37

Installing the Piezometer.

37

Flushing out the Piezometer.

39

10
11
12
13
14
15
16

17

Gardner (38).

Applying Suction to the Piezometer.

Stopper is for a 2-lnch Piezometer.

Upper
39

Measuring Water Elevation with Reel-Type Electric

Probe

44

Removing a l-inch Piezometer with Veihmeyer

Soil Tube Jack.

44

Cirmiit for Location of Equipotential Surfaces.

frevert (27).

47

Circuit for Determination of A-function* Frevert

(27)

47

Plot of Field Data to Show Linear Relationship

Between In(d-y) and Time t.

48

Variatioa of A-funotion with Diameter of Cavity for

a 4-inch Long Cavity (For Values of s and d See
Text).
Variation of A-function with Length of Cavity for
a 1-inch Diameter Cavity (For Values of s and d
See Text).

69

69

-riFig.
18

Page
fariation of A-fuactioii ?dth. s, for a Cavity 4iaehes leag and l-iach in Maaeter.

71

19

Iqiiipotential Surfaces for a Piezometer Cavity.

71

20

Sxample of Mtbaiaim Proeedure,

75

21

Iquipoteatial Plots for Gase of iDralnag of

Uniform Soil. S Text for Details.

80

22

Point lar a Gurved Boundary.

81

23

Point on an Interface.-

86

24

Point at Upper Oorner.

See Text.

89

25

Point at Lower Oornr,

S@e Text.

89

26

Iquipottntial Plot for Drainage of a Two-layered

Soil.

91

Icpipotential Plot for Drainage of a Two-layered

Soil.

93

27

28 fiquipotential
Soil

Garves for Drainage of a Two-layered

93

I.

IITROOTGTIOir

It has long toen recognized that soil permeability is an

iBEportant factor in land management, soil conservation and land
drainage.

Muoh work has been dxjne on methods of measuring soil

permeability and various sehemes have been devised to apply the

information to the design of drainage systems, dams and other
engineering structures.
In general, the measurement of soil permeability has con
sisted of taking a sample of soil from the field into the labo
ratory in either a disturbed or an undisturbed state, passing
water through the sample, and determining its permeability by
use of suitable equations.

Such methods of permeability meas

urement have been unsatisfactory for many reasons.

Recently several faethods of determining soil permeability
in the field have been proposed.

In 1936 Hooghoudt augered out

a hole in the soil below the water table and observed the rate
of rise of water in the hole.

By means of approxiLiate formu

las he was able to calculate the soil permeability.

Kirkham

and Van Bavel pointed out certain defects in Hooghoudt's formu

la and derived a more exact equation based on a solution of
Laplace's equation.

They tried out the method \?ith success on

several Iowa soils.

In 1945, Kirkham proposed a field method of measuring
soil permeability x'yhich consisted of driving pipes into the

-2soil below a water table and meaiuring the outflow of water

frotB the pipe into the soil.
out by Fre^ert and Kirfchaa,

This proposed method was tried

They found that it was necessary

to remote the soil from the interior of the pipe and they also
noted that a lauoh laore aecurate measure of the permeability was
obtained by permitting the water to flow into the pipe from the
soil rather than using the outward flow from the pipe into the
soil.

Because driving the pipes into the soil compresaed the

soil, they had to use pipes of large diaaeter (8 inches).

The

depths to which the soil permeability could be measured were

lifflited to 36 inches because of the equipment used, and special
tools were needed to remove the pipes from the ground.
the preceding field methods have in common certain ad
vantages over any laboratory methods which have been proposed.
First, the soil permeability is measured in situ; second, soil
water is itself used for the measurement; and third, root holes,
worm holes, and rocks have a negligible effect on the deter
mination.
There are, however, certain limitations to these field
aethods.

Both of them are limited to shallow depths below the

soil surface and the auger hole method, as it gives a sort of

average permeability over the length of the auger hole, will
not Indicate the permeability of specific layers or horizons.
To overcome the limitations of the above methods, a new
procedure, utilizing pipes of small diameter (piezometers).

-o
has been developed.

Tiie method consists basically in measuring

the ratt of flow into a canity at the base of an emptied pie^offleter.

Advantages of the method are (l) the permeability can

be measured to great depths, (2) the permeability of any layer

in the soil can be measured, (3) the method is quick, acciirate
and simple,

fhe deirelopent of this method, \i^ich includes

use of a three-dimensloiial electric analogue of the groundwater

problem., constitutes two parts of t&ls thesis.
Kirkhaa has solTed the problem of steady-state flow into
tile drains in a uniform soil but there is no general solution
that will handle all of the complex Tarlations found in natural
soil.

Similar types of problems are encountered in the fields

of heat flow, electricity, and hydrodynamics and various in

vestigators have developed numerical methods of obtaining ap
proximate solutions to any desired degree of accuracy.

Because

of the similarity of these types of problems, it is evident

that numerical method qbm also be applied to problems of land
drainage.

The development and application of suitable numerical

methods to land drainage constitute a third part of this thesis.

-4-

II. REVIEl OF IiITSRATURE

A fflore thorough review of the literature on soil permeabllitir can he found in Freveart's (27) thesis.

However this

subject will be covered here in brief.

A.

Darcy's Law

The rBDvemeat of fluids through porous media has long been

of great practical importance to agriculturists concerned with
irrigation and drainage and to engineers interested in the flow
of fluids into wells and through filter beds.

Since most porous

media can be regarded as a macroscopic collection of more or

less discontimioua capillary tubes, the first experimental
studies which can b regarded as forming the basis of our present-day knowledge of water flow through soils were performed
by Hagen (32) in 1839 and by Poiseuille (53) in 1846.

These

investigators studied the flow of fluids through capillary

tubes and from their observationa concluded that the rate of
flow is

proportional 1;o the hydraulic gradient.

In 1846 a French hydraulic eiigineer, Darcy, used experi

mental methods to study the flow of water thro\;^h sand filters.
His classic experiments led to the result - now referred to as
Darcy*9 law - that the rate of flow of water through the fil
ter bed was directly proportional to the area of the sand and
to the difference of the fluid heads at the inlet and outlet

5
faces of thte bed, and In-^erssly proportional to the thickness
of the hed.

(See fig. 1.) Expressed mathematically, Darcy*8

law becomes
Q

K I A

(1)

where
Q = discharge "velocity
K = coeffioient of periseatoility
A s cross sectional area of bed
L = length of bed
h difference in head between outlet and inlet faces.
The B'alidity of Darcy*s law has been confirmed by many
experimenters, aost of ^om used sand separates in their tests.
Since OUT main interest is the application of Darcy*3 law to
the moveraent of water through soils we may consider SeelhetM
(60) to be the first in-s'estigator to use soil in a check of
Darcy's law.

The data (in part) -hich SeeHieim obtained is

Included in table 1.
fable 1
Relation of Pressure to Flow Through Clay - Seelheim (60)
Pressure
150 Gffl.
100 cm.

list
60
60

SjISas.
12 c
13 c

Me^
0.59 cc
0.39 cc

fhese results, though limited, ooafirm the application of

Darey*s latt to fluid flow through clay.
Teraa^hi (66) in 1925 showed that Darcy*s law can be ap
plied to the flow of water through clay and concluded that it

y=k(h,-h;)

h,-h.

t-v=kV$]
V

Fig. 1.

Illustrating Darcy*s Law.

Gardner (38)

^7may be safely applied to grouad-irater flow.

Kiag (40) ia a review of earlier work pointed out that the

relatioaship between the quantity of water transmitted throvigh
the soil with increasing pressure (or hydraulic gradient) did
not always follow a straight line.

In many cases the quantity

of water discharged did not increase as much as the pressiire.

This was especially true for very porous substances such as
gravel or where the hydraulic gradient was so large as to
cause high velocities of the water moving through the material.
This departure from Darcy's law has been satisfactorily
explained on the basis of Reynolds*(55) work with capillary
tubes.

At low velocities the movement of the water is laminar

(in straight lines) and the only force opposing its motion is
the resistance of the walls of the capillary tubes (or soil
parti'cles).

Ihen its velocity increases over a certain value,

then the flow becoaies turbulent and part of the energy of the
moving liquid is dissipated in eddy losses.

The flow is re

duced because of these energy losses in eddy currents (see

Muskat 51, p. 56).

The velocities normally encoimtered in

the aoveiaent of water through soils are well within the range
of the validity of Darcy's law and so we may safely apply it
in our soils work.

B.

Equation of Flow

Darcy's law as originally formulated applied to flow in

^8one direction and nay be rewritten ia differential form, substi

tuting Telooity v for discharge Q, and taking A 1 as
V = K

(2)
oh

where

is the hydraulic gradient in the direction h.

Slichter (63) was the first to show how Darey's law can be ap
plied to flow in any direction by rewriting eq, (3) in the form

where 4^
os

hydraulic gradient la any direction s.

In case no

external attractive forces are acting and K^, Ky, K^, represent
the soil peraeability in the x, y, z directions then the equa
tion of continuity as deriTed by Slichter (63) ia
(3)
which is a osathematical statement that the liquid is inco/apreasible - that is, a given taass of the liquid does not chaise
its volume during the given motion.
If Kjj s Ky - Kg (ie, the soil is hostogenecaas) then eq. (3)
becomes
^ ^

ill = 0

which is Laplace's equation, an equation that occurs frequently

in mathematical physios.
function.

The function

is called a potential

Thus the solution of any ground water problem depends

on a solution of Laplace's equation as pointed out by Slichter

(63).
To Slichter also goes credit for showing that the motion

^9f ground water is analogous to the flow of electricity, the

flw of heat, or to a problem In the steady motion of a perfect
fluid.

Slichter noted that the gravity factor should be included

in the equations of motion and much of the present day work on

the theory of ground-water moveffient is based on the relationships
discoirered toy him.

C.

Permeability Onita

Although soil permeability has been under intensive investi

gation for almost 100 years, there is still no generally accepted
method of expressing the permeatoillty factor or coefficient (K in
eq. 1).

Richards (56) suggested that k in the equation

V s ki

(5)

where v quantity of water flowing per unit time per unit area
i as hydraulic grMient
be used.

This is the Daroy coefficient and has the convenient

dimensions of velocity (e.g. c per sec).

(The effect of vis

cosity is included in eq. (S) as a dimensionless ratio.)

Veloc

ity of flow is in common usage in engineering and hydrologic

work and possesses certain advantage over other methods of ex
pression.
BodraanC?) has suggested a unit of permeability having the
dimension of time according to the following equation.
^8 = f-i-l/h
where
Pg permeability coefficient

-10g/h = hydraulic gradient

Q/t = discharge velocity
A orosB sectional area
1 length of column
Wyckoff, Botsetj Muskat and leed (73) have introduced th
visoosity of the fluid into Darcy's equation, thus giving K, tie
dimension of an area.

This coefficient waa called the "darcy*

and was defined as the voluiae of fluid of one centipoise vis

oosity passing throu^ a one square om. cross section in one
second under the action of a pressure gradient of one atmos
phere per oai.
Richards (57) later presented a new unit called the "dar**,
which had the dimension of time and was numerically equal to
the mass of the liquid crossing unit area in unit time per
unit pressure gradient per centipoise viscosity*

This uait

had the advantage that it was independent of the system of

measurements used (metric or English) and was also adaptable
to a wide range in permeabilities.
Edlefsen (20, p. 431* 432) pointed out that "In some
studies, it is probably preferable to use one form of the
permeability-coefficient, while in other studies, it might
be more convenient to use a different form."

D.

Methods of Measurement

lo attempt will be made here to review all of the various

methods of measuring permeability, sinoe there are almost as

many methods in use as there are investigators, tout rather some
of the general features of each method will be described.
1.

Laboratory methods
a.

The disturbed sample.

In the use of disturbed soil samples the soil ia collected

in the field and is packed into permeameter tubes in the labo
ratory.

Mo special attempt is laade to preserve the natural

structure of the soil when filling the permeameter.

This

method has met with success in its application on the struc

tureless soils of Western United States in studies of the
effects of various treatments on soil permeability.

The

permeability values obtained are not necessarily related to

the permeability of the soil in th^ field but serve to indi
cate the relative effects that various soil treatments might
have in the field.

Fireman (22, p. 337) discussed this method

in detail and stated that,

In many cases permeability values obtained In
the laboratory may not even sp proximate the
percolation rates .... Preliminary tests
indicate that, regardless of the correlation
between laboratory and field percolation rates,
the relative change in permeability obtained in
the laboratory as a result of any given treat
ment is closely correlated with the relative
change in percolation rate obtained in the
field as a result of a similar treatment.
b.

The undist\irbed sample.

arious investigators have devised methods of obtaining

so-called undisturbed soil samples.

In general these methods

h&we consisted of inserting a metal cylinder into the soil and

removing the cylinier filled itfith soil.

The soil-filled cylin

der is then taken into the laboratory where permeability measmrements are made.

Large errors enter into the determination

because of the presence of root holes and rock in a sample that

is necessarily small and the tremendous variation between repli
cates makes it impractical to apply the results in the field.
2.

Indirect methods
Many soil properties such as hygroscopicity (59), pore

sise distribution (3, 4, 5, 49, 52), point of inflection on

a pF curve (3), size and shape of the soil grains (63), per
centage of elutriable particles of certain sizes (59), have
been used in an effort to find some soil property that can be
correlated with soil permeability.

In spite of the great amount

of work done no generally satisfactory method has been foxuid

that can be applied to soils over a wide region.

Firensan (23,

p. 340) in a discussion of indirect methods of making permeability

measurements concluded that,
Kiese methods have not been of particular value
except in special cases such as the evaluation
of probable permeability to air. . .
.,
nor do they usually involve a saving in effort
or expense.
3.

Field methods
a.

Unllned wells.

fenzel (71) in a review of the literature on methods of

permeability measurements classified the methods of measuring

-13permeaMllty froa unlined wells as those imrolvliig observations

oa the drawdown of the well and those involving the rate of re^
covery of the water table.

?ariome equations have been devel

oped for calculation of the permeability from the above meas

urements.

A basic assumption of the derivation is that the

Gone of depression around the discharging well has reached

equilibrium (steady-state flow of water).

Jacob (39) has de

veloped a foriTOla based on a nonequilibriiiffl state.

In the field of soils Diserens (18) and Kozcny (47) have
proposed digging an auger hole into the soil, and observing the
rate of rise of the water in the hole.

Hooghoudt (38) developed

formulas for calculating the soil permeability from the observed

rate of rise of the water in the hole but his equations are not
based on solutions of Laplace*s equation and their validity has
been questioned by Kirkham (45).

Kirlcham and Van Bavel (45) and

Tan Bavel and Klrkham (68) have developed ore exact formulas
based on solutions of Laplace*s equation and have applied the
method to several Iowa soils.
b.

Lined wells.

The lined well method consists of inserting a tube, pipe,

or cylinder into the soil and observing the rate of flow of
water into or out of the pipe.

This method has been widely

used (26) in a study of the relative rates of infiltration of

water into different soils.

Suitable equations for calculating

the soil perioeability from these observations were lacking un

til lirkham (41) in 1945 solved the problem for a steady-state

14oondition.

The method of permeaMllty meaeureraent developed

in this paper makes use of the formulas given by Kirkham and

follows somewhat the methods proposed by him.
larlier in 1932 Kozmy (46) proposed a similar method of
driving tubes 10 Offl in diameter below the water table and ob
serving the rate of fall of wster in the pipe after the soil had
been removed from the interior of the pipe.

The formulas which

Kozmy developed were based on several approximations and

Kirkhaffi (41) has questioned their validity,
e.

Tracers.

Tracers have been effectively used to measure the velocity

of ground-water movement and the results have been summarized
by Tolman (67) and fensel (71).

Tracers that have been used

include salt solutions of 11401 and laOlt and dyes such as

flourescein.
The movement of the salt solutions has been followed by
observing the change in electrical conductivity of the ground
water.

Wenzal (71) in a discussion of the use of salt solu

tions as tracers pointed out that salt water is heavier than

the ground-water and laay flow downward so as to miss the point
at which the sample is taken,
Tolman (67) in a discussion of the use of dyes aa tracers
stated that they may be more accurate than salt solutions but
their us is Halted due to their interaction with the organic
matter of the soil.

15-

S.

Factors Affecting Soil Permeability

learly all of the work that has been done on the factors
which affect soil permeability has been done in the laboratory
and most of the results obtained are merely qualitative in char
acter.

Much work remains to be done to determine the magnitudes

of the effects that these 'Various factors have on soil permea

bility in the field.
!

Intrap-ped air
A review of the literature indicates that air entrapped in

the pores of the soil is one of the ma^or causes for the failure
of permeability neasuretaents made in the laboratory.

Evidence

is lacking on the effect in the field of air entrapped in the

soil but judging froa laboratory experiments it must be of a
considerable magnitude.

Its influence must be particularly

great during the initial wetting of a dry soil by rain or by

irrigation water.

As the air is dissolved in the percolating

water, its effect on the permeability is decreased until after

a period of tlrae all of the air has been dissolved out of the
soil.
Ohristianacn (14) on the basis of his work on the effect
of entrapped air, concluded that the permeability of the soil
may be reduced to 1/30 of its air-free permeability.

The time

required to dissolve the air out of the soil in the percolate

varied from 3 to 53 days.

In laboratory permeability determinations, it was at first

tbought that the air could b driten out of the soil by wetting
the soil froffl the bottom.

This idea is now regarded as falla-

eious since it really makes little difference whether the soil

is wetted from the bottoa or the top.

The qualitative reason

for the effect of the entrapped air reraains the same.

According

to Wyckoff and Botset (72) an explanation of this phenomena was

given by Jamia for the flow of fluids in constricted capillary
tubes that contained gas bubbles.

The presence of a constric

tion or an abrupt change in diameter of the capillary acts as

an obstruction to the gas bubble and it will not flow through
the non-uniform section unless a certain critical force is
exerted on it.

A bubble advancing into the constricted portion

of the tube must suffer distortion which Involves an increase

in the surface energy at the Interface between the gas and the
liquid at the small radius end of the bubble.

This makes the

surface tension forces unequal with the net stress being to

drive the bubble back and out of the constriction.

The appli

cation of a slight external pressure will only serve to drive

the bubble back into the constriction where it will remain
stationary in equilibriuM with the externally applied force.
If this qualitative explanation is valid then it does liot make
any difference ?jhether the water is introduced from the top or
from the bottom as has been advocated by some investigators.
Since air is soluble in water and will in time be dis
solved in the water percolating through the soil, a novel

-17method of eliminating the air from the soil was suggested by

ChTistismsen, Fireman, and Allison (16), who first displaced
the soil air with carbon dioxide.

Carbon dioxide is readily

soluble in water and when water is added to the soil the car
bon dioxide goes into solution and an air-free soil results.
Wetting the soil under a mcuuffi is another technique used
to eliminate the errors due to entrapped air (64).
2.

Microorganisms
Recent investigation on the influence of microorganisms

on soil structure and soil permeability indicate that they may

be much more important in their effect on soil permeability than

heretofore realized.
Although ffork by early investigators (69) indicated that
Microbes and their decomposition products may influence soil
perraeability, Allison (l) in 1947 was the first to show con
clusively the magnitude of this effect.

His results show that

under conditions of long submergence the microbial bodies and

the gxjuas and slimes produced in the decomposition of organic

matter may plug up the pores of the soil so that the permea
bility is markedly reduced.

The generalized perraeability

curve obtained in the laboratory under long submergence is

shown in fig. 3.

Allison (l, p. 440, 441) gives the following

explanation of this curve.

Phase 1. After initiating field or laboratory
tests, the permeability decreases to a minimum.
On highly permeable soils this initial decrease
is small, or nonexistent, but for relatively
impermeable soils, it may be appreciable and

-18-

FIG. 2 , CHANGES IN PERMEABILITY OF SOILS DURING LONG SUBMERGENCK""-^H i S Oil

PCRMEABtLITY
CW/HR.

STERILe SOIL a WATER

STERILE SOIL-REINOCULATEO

CONTROL

20

30

40

50

60

70

TIME-DAYS

FIG.3. PERMEABILITY-TIME CURVES FOR HANFORD LOAM

U NDER P ROLONGED S UBMERGENCE AXLISOH

-19-
OQatinme for 10 to 20 days before the second
phase of Increase 1 apparent. The decrease
in permeability it probably due to structural
changes resulting in part from swelling and
dispersion f the dry soil upon wetting and
in part to dispersion resulting fro a de
crease in electrolyte content of the soil
solution as any salts present are removed
in the percolate.

Phase 2. When soils are wetted from the

surface dowiirard, considerable air Is en
trapped in the pores (l, 10). As the air
is dissolved and removed in the percolating
water, the permeability gradually increases
attaining a oaxiiBUffl ^en all or nearly all
of the entrapped air is removed, the mini
mum peraeabllity appears to be the resultant
of two opposing phenomena, that is, the
forces described for Phase 1 tending to re
duce permeability from the beginning and the
forces of air reoved tending to increase
permeability.
Phase 3. After the maximim is reached, the
perffleablllty decreases with time, rather
rapidly at first then more slowly until after
8 to 4 weeks the rate is usually a small por
tion of its original value. Frequently it
has been observed in laboratory tests that
the maxiiMUBi permeability is reached before
the last of the entrapped air is removed.
The gradual sealing of the soil during the third phase is
due (as Allison has proven) to biological clogging of the soil
pores with microbial cells and their synthesized products,
slimes, or polysaccharides.
Fig. S shows the results that Allison obtained by using a
soli sterilissed with ethylene oxide gas as contrasted to an
unsterilized soil and a sterile soll-relnoculated.

The ster

ilised soils reach a constant high permeability value whereas

the unsterilized soil and the sterile soll-relnoculated have
decreasing permeability with time due to the microbial action

20""
mnder long ambaergence.
3.

Salts
All soils contain oolloidal Material and the degree of

dispersion of the oolloidal material has an effect on the soil

permeability.

This effect is particularly ia5)ortant in the

structureless soils of the fleet (33, S4) and of considerably

leas importance in well-g^gregated prairie soils of the Mid
feet.
Harris (35) in a series of classic experiments on the
perraeability of alkali soils fomd that the presence of sodium
on the base exchaage co^lex caused a decrease in the soil
permeability.

The decrease was attributed to an increase in

the dispersion of the colloidal particles due to the presence

of sodium in the base exchange complex.
In the reclamation of alkali lands it has been observed
that there is an initial decrease in the soil perraeability
when the excess salts are washed out of the soil.

The presence

of excess salts represses the dispersion of the colloids.

When

these salts are washed out the soil is deflocculated and the
permeability decreases.

In order to flocculate the soil again

it ia necessary to replace the sodium in the base exchange

complex with calcixiffl ?/hieh reflocculates the soil.
4.

Temperature
Poiseuille (53) studied the influence of temperature on

Tiscosity of water and determined the relationship between them.

-31^
Hagn (1869) irerifiad Poiseuille^s relationship and found a 3^
increase in peraeaMlity for every increase in temperature by
one degree.
Gustafsson (31) pointed out that as long as Darcy*s law
holds (laminar flow) there is a linear relationship between
teaperature and perraeatoility.

In the region where Barcy*s law

is Talid only the forces of friation oppose the movement of the

water and these frictional forces are directly proportional to
the viscosity of the water.
As an example of the magnitude of the effect of temperature,
in the spring of the year with water at 1 C as compared to sum
mer at 23 G the permeability ratios are 100:163 due to temper
ature.

However Duley and Domingo (19) in field tests did not

find any practical signifioaace in the effect of temperature

variations on soil peraatoility,

F.
1.

Solution of Flow Probleais

Analytical methods
a.

Dupult-yorohheimer theory

Although the assuaptions of the Dupuit-Forchheimer theory

of gravity-flow systems have been shown by Muskat

(51, p. 359)

and others (fO) to be so questionable as to make the whole theory

untrustworthy, its widespread use even today makes it necessary
to consider it in some detail.

(It should be pointed out that

although the theory is based on erroneous ass\nptions the results

-22-

that are obtaiaed are often surprisingly close to those given

empirically or by exact calculation.)
Dupuit assumed that for small inclinations of the free
surface of a gravity-flow system the streamlines can be taken
as horizontal, and are to toe associated with velocities which
are proportional to the slope of the free surface, but arc in
dependent of the depth.

la other words, all the flow is in a

horizontal direction and the rate of flow depends on the slope

of the water table,

fhis i of course contrary to the flow

patterns obtained in sand-tank experiments (34) and electrical

analogues (10) and analytical solutions (44), which have shown
the true, circuituous paths taken by water particles.

The

theory thus assuaes that a system will have no gravity-flow

characteristics which is entirely contradictory to the impli
cations of Barcy's law as pointed out by Muskat (51).
lumeroua investigators (2, 17, 46, 47) have applied the
Dupuit-Forchheimer theory to the problem of land drainage and
although the aathematical devices used to get their results
have differed, the equations derived are all very similar.
Russell (59) has discussed the literature on the subject be
fore 1934 and the following example will largely follow his
text.

Certain assuiBptions are aiade: (l) the soil is homog

eneous; (2) rain water percolates uniformly through the soil

till it reaches the water table which starts rising; (3) when
the ground water reaches above the drains, it will start to

~.S.3-

flow lato them, the rate of movement being greatest in their

iramediat# neighborhood and slowest halfway between the drains;
(4) the me&n horizontal oomponent, v, of the rate of flow
through any cross section PQ is proportional to the slope,
of the groimd-water aurfaee at P, and, {5) no \mter moveQX

ffleat oijcurs through the subsoil below the drain level AB.
The derivation follows: (See fig. 4)
* fed.y
dx
Ci a ky^
dx
Sinee the ground water surface reaains stationary, Q is equal
to the amount of water percolating downwards from the soil sur
face In unit time between P and 0, therefore
Q at q(^ - x)

where q is the afflouat of water percolating in unit time per

unit area of soil surface, therefore

and on integration
ic.y^ a EX - #
i
This ie the equation of an ellipse with center at D and semiaxis 1 and I'l
fhus the sumffiit of the ground-water surface
1
reaches a maximm height h; where h
Hence if the drains
are laid at a depth

the ground water will not rise above a

depth h* below the surface where h* s d-h.

G R O U N D

Free Water

S U R F A C E

Surface

(Phreatic Surface)

DRAIN
TUBE

DRAIN
TUBE
E

Fig. 4.

Dupuit-Forchheimer Theory of Soil Drainage

The equation may b@ re-writtern

d - h- = I
K is determinable and bene# if q (the maxin^ra rate of influx
of water whieM the system will be required to deal with) is
known, the depth and spacing can be chosen in such a way that
the ground water will neirer rise nearer to the surface than
soiae specified distance h, hieh must be sufficiently great
to present damage from water-logging.
b.

Kirtehaa-Gardner approach^

ly assuming a steady state condition with the potential

function (or its aormal derimtive) known over the boundaries
of the region considered, it is possible to get solutions of
liaplaee*8 equation that will satisfy the known bomdary con
ditions.

These solutions can be used (after considerable

mathematical manipulation) to determine the flow into drain

tubes embedded in the soil,

Slichter (63) first pointed out

the possibilities of this approach and Gardner et al (39) made

application of it to the problem of artesian flow.

Recently

Kirkham (43) has solved the problem of flow into a series of

drain tubes embedded in a hotaogeneous soil.

In addition he

has solved the problem of flow into drain tubes embedded in a

two-layered soil (4(4), each layer having a different permea
bility.

These solutions were obtained using the method of

images (See Muskat 51, p. 175) and the results were expressed
in conjugate fmctions.

The real and imaginary parts were

-26-

separated to give a potential function ajid a stream function.

The normal derimtlre of the potential was obtained at the soil

surface and this aortaal derlirative was then integrated over the
surface of the soil.

Multiplication by the soil permeability

gave the quantity of water flowing across the surface of the

soil and therefor the quantity of water flowing into the drain
tubes.
In this analysis the assumption was made that the surface
of the soil was everywhere at the saae potential (ie. ponded
water).

The falling water table was not taken into accoiint in

lirkha*s. analysis,
2.

Qraohical method of Daohltr.. Gasagrande. and Forchheliaer

Muskat (51) la a discussion of Daohler^s work, described

the application of the graphical method to seepage through

dams made of permeable material,

fhe Initial problem is the

location of the surface of seepage.

The Inflow face, main body,

and outflow face of the dam are treated as separate flow sys
tems by different approximations, and are then synthesized by
the requirements that the fluxes through each shoxild be equal,
and that the fluid heads should be continuous in passing from
on part to the other.

4fter the surface of seepage has been

located, the equipotential lines, each one representing a constiunt fraction of the total loss in head h, are drawn.

Streanh-

linea are chosen to be orthogonal to the eqiAipotential so that

the saae fraction

of the total seepage passes between any

-27pair of neiglibo3rlag flow lines.

of a series of squares.

The resulting net will consist

If one aueoeeds in plotting two sets

of curTea so that they intersect at right angles, forraing

squares and fulfilling tbe boundary condition, then one has

obtained a graphical solution of the problem.

Oaaagrande (9)

applied the method to soil which are anisotropic with regard

to permeability.

He showed that all dimensions in the direction

k max are reduced by the factor

/KsIb or all dimensions in the

V k max
j

direction of t mlm. are increased by the factor Vk^n*

A

probleffl is solved for the case of a soil with uniform permea

bility and then the above conditions are applied.

The average

permeability will be
* Vk in k max
and the flow will b

q t
where

is the directional derivative of the potential.

The answer obtained by use of the graphical method is

only approximate but is of the correct order of magnitude and

use has been made of the method in the design of earth dams.
3.

fhe hodograah
A hodograph is a representation of a dynamical system

In which the coordinates are the velocity components of the

particles of the system.

The method of treatment is difficult.

A description by Muskat (51) will be followed here.

Although

the method was developed earlier by Helmholtz and Kirchoff,

-28-

Haael (33) first made considerable application of it to t^odimensional seepage systeffls,

fhe systems included simulta-

neoualy ii^ermeatole boundaries, constant-potential surfaces,

and surfaces of seepage,

fhe liodograph of a flow line is the

cunre which one obtains when plotting from one origin velocity
vectors for all the points of the flow line,

^e straight line

Gonneeting the origin with on point on the hodograph represents

the fflagnitude and direction of the velocity for the correspond
ing point in the flow line.
Since the velocity along the free water surface is propor
tional to the Bine of the slope, the hodograph for the line of
seepage is a circle with diameter equal to the coefficient of
permeability.

The hodograph for a straight boundary is a

straight line.
Once the boundaries of the system are fixed (in the
hodograph plane) the methods of conjugate-function theory can
be applied to the final solution of the problem, although the
transformations of the circular segments representing the free
surfaces involve the theory of modular elliptic functions (See
Musfcat 51, p. 301).
4.

Mesi^rane aaalomie
la a study of the uplift pressures on large dams, Brahtz

(8) developed a ajembrane analogue to obtain solutions for twodiaiensional probleffls*

Bie analogue takes advantage of the fact

that the differential equation for the steady state potentials

29-,
for the percolation witMn an earth or concrete mass has the
same form as the differential eqmatioa for small normal defleotiona of a niforly stretched rubber mefflbrane (Laplace's equatioa).

The relative ordinates at all points along the bound

aries of the membrane are made proportional to the boundary

pressures at corresponding points in the structure.

The or

dinates at all interior points will then be proportional to

the pressures at corresponding interior points in the prototype.

The analogue is set up to represent the field conditions

and the ordinates of the rubber membrane are measured by meajas

of a micrcmeter device.
The asufflptions Bde in the use of the meiabrane analogue
are (1) the materials of the d&in and the foundation are honKJgeneoms (2) the merabrane i of infinite extent in all directions
(3) the membrane is only slightly displaced*
llestrlcal analogue
The electrical method of flow analysis in seepage problems
was first proposed by Pavlovsky in 1932 [^Bee Low, Dams, Natl,
les. Com, 1938 Wash. D.O., for a discussion of the electrical
method] and is based on the relationship between Ohm's law and
Darcy*e law.

Ohffi*s law, which expresses the fundamental rela

tion for the flow of an electric current, is expressed by the

equation
i a - o-

(6)

where i s curreat per imit area

speoiflc oonductivlty
voltage gradient in the direetion s.

The minus siga

indicates that the potential decreases as the current progresses

%n the positive direction.

Darcy*s law for the flow through

porous media is
Q

|4.

(7)

OS

where Q quantity of water flowing per unit area

k permeability (or conductivity)
hydraulic gradient in direction s
Equation (6) is identical with equation (7).

Since the princi

ples of flow are similar, with the saaie conditions as regards

pressure and path of flow, the flow itself will be similar.
The electrical analogue can be used to trace both the
flow net and the equipotential net and the resulting figures
can be used to coapmte the quantity of flow throiigh the sec
tion investigated*
Ohilds {10 11, 12, 13) has made the most intensive appli
cation of the electrical analogue method to the study of land
drainage.

He soaked sheets of filter paper in colloidal graph-

it and reproduced a section of the field case.

Tsa used to represent the drain tube.

A copper foil

The vertical flow in the

soil above the water table was represented by current input

between a series of strips conducting material insulated from
each other.

-31-.
6..

Muwerlcal methods
Iimcrical methods of solving two-dimeasional partial

differential equations involve the replacement of the continmum of points in a region considered, by a discrete set of
points, and the replacement of the differential equation by a
finite difference equation.

Llebmaan (48) was the first to

show how to replace Laplace*s equation by a finite difference

approximation and obtain a solution by an iterative procedure.
In the process developed by Liebaaan, a square net of points
is laid over the region and approximate values (best guesses)
of the function are assigned to interior points while known
values of the function are placed on the boundary points.

The

net is then repeatedly traversed, replacing the value at each

interior point by the mean of the values at the four neighboring
points, using the new values in the improvement of the succeeding
points.

After a number of traverses, the function at each point

will converge to a solution of the finite difference equation.

The accuracy of the answer obtained and the rate of convergence
-will both depend on the size of the net spacing.
The difference between the solution of the approximating
difference eqpiation and the differential equation representing
the true solution has been investigated by Richardson (58),
showed that on a net of interval h the difference is of the form
A (X,y)h^+B(x,y)h^4-C(x,y) h+
Since only even powers of h enter here, the difference solution
approaches the differential solution rapidly as h approaches

332ro,

As pointed out toy Shortly et al (62), if h is suffi-

0iBtly siaall the error is proportional to

and by making

two solmtions with different h, one can estimate the differ#ae of @a(^ from the solution of the continuous problem,
Shortley, Isller, and Fried (62) have studied the rate
of convergence of the Lietomann procedure by investigating the
rate of convergence of an arbitrary *error" function as the
saffle iaprovement formla is used repeatedly.

The "error*

function has zero botindary values and converges to zero every

where in the region,

the rates of convergence of the "error*

function and the true function are the same.

As a result of

their investigations they conclude that the rapidity of con

vergence varies inversely as the number of points in the square
region under consideration,

for example, the error at any point

in a square region with 81 interior points is reduced to about

nine-tenths of itself by each traverse after it has been im
proved a few times.

At this rate, it would take about 23

traverse to reduce the error to one-tenth of its initial

value.

For a net of 389 interior points, about 75 traverses

would be necessary.
In an effort to speed up the rate of convergence, various
procedures have been devised hich treat whole groups of points.
This so called block procedure has been applied by Shortley
(62), Southwell (65), and others with success.
Southwell (65) developed the "relaxation* method which is

"33"
superior to the Itiebaiaaa proeedure in the length of tirae re
quired to solire th protolem.

fhe relaxation method was devel

oped froBi a consideration of statics problems and Southwell

generally speaks in termt of a tension net as an approximation
to a soap film or membrane.

A simple explanation of the relax

ation method has been given by Imroons (21).

In the relaxation method, instead of dealing with the
values of the wanted ftinetion, (j> , at the interior points the
residuals are computed and recorded.

Referring to fig. 5 the

residual R is coE^uted by the foranila (the prime indicates that

the value is an approximation)
R 4>' (i>^ -H

+ <i>^' H <i>^

fhe R thus coi^uted can be thought of as an interior sink which

aittSt be reaoved.

If 0' is altered at any one point, there will

be a change in the residuals at each of the four points sur

rounding

d)' .

Specifically, if

is altered by -4 units each

of the residuals at th four surrounding points will increase

by one unit,

lach change of a <P' at an interior point will

cause a redistribution of the residuals according to the

"relaxation" pattern of fig. 6.

Hie calculator adjusts

and the H*s until the residuals are sufficiently small.

4'
The

problem is then considered solved.

An interesting application of the relaxation method has
been made by <*illes (SO) to problems involving two dependent
variables.

Since the ecpiipotential curves and the streamline

curves are conjugate plane-harmonic functions, they are con-

4),

^4

(I)o

(|)2.

*^3

fiit S.

fif# i#

fnflwtii

3&~'

nscted toy the Gauehy-Rieffianm eqmatioas.

5 0 ^ a tp
Sx ' c)^

'

^
^ ^

^ yy

Gilles developed suitable equation so that the two wanted func

tions,

f &mi <p can b deterffiined siaultaneoualy.

However in

his words, *The process is both laborious and tedious*.

Moskovitz (50) has presented a process which yields pre
cisely the convergent values of the function obtained by in
finitely Biany traverses of the region,

fhe Liebmaim formula

is used for several traverses; then, succeeding values of the

improved function are caloulsted by aeaas of tables included in
Mo8kovitE*s paper.

Only regions with rectangular boundaries

can be treated by this aethod.

Shaw and Southwell (61) have applied the relaxation method
to probleffis of percolation under a da and have treated the case
of flow thro^u^ layered porous material having different permea
bilities in each layer.
Fox (25) has shows how values of the function may be com
puted at points close to an irregular boundary.

-36-

III.

fHE PROBLEM

The problem can be divided Into three related perta,

(l) the dev#lpffint of a field procedure of measuring soil
peraeability beneath a water table, (2) the use of the threedimeasional electrical analogue for the determination of cer
tain constants in the perffleability equation and the use of the
electrical analogue in the study of soil factors affecting
flow, (3) the application of peraeability data to a drainage
problem solved by numerical aethoda,

A.
1*

PieEometer Method of Measuring Soil Permeability

pmcedur#

The field procedure as finally developed is as follows: ^

The surface sod is removed from the soil and a hole is augered
out to a depth of about 6 inches below the surface of the soil.
The auger used is one of 1/16 inch smaller diameter than the
inside of the piezoetr.

A piezometer (which is an unperfo-

rated pipe) is then driven about 5 inches into the augerd-out

hole. ,(One inch thia-walled electrical conduit has proven
satisfactory for use as a piezsmeter.)

The auger is inserted

inside of the piezometer and a cavity is augered out for another

six inches below the end of the piezometer which is then drivea
with light taps of the aaul (fig. 8) into the soil for another

-37-

^ PIEZOMETER
SOD REMOVED

^SOIL SURFACE

\i n 11 I / I ! I I ! II III I
WATER TABLE

-2R
y

I^CAVITY

/ FLOW
LINE

s 2r
IMPERVIOUS
I I I I I I / I I I I I I 11
111111111111
111 11 I I / i/iLIAYFR

7.

Pl@zQt? Met&od of Mea8iufmat of

Permeatolllty.

8.

lastailing th Piezometer.

38*"

5 inoliea,

Tkis procedure of successive augerings -Is repated

until tlie piezoaeter is at the desired depth below the surface

of the soil.

It should be emphasized that the piezometer is

always forced into a cavity with diameter slightly smaller

than that of the inside diameter of the piezometer.

It is

never forced into natural soil without augering.

Ihen the piezometer is at the desired depth a cavity of
any convenient length is augered out beneath the piezometer.
Water will flow into this cavity during the permeability test
and it is essential to auger it out with care.
fhen the piezometer ha been placed at the proper depth,
a hose from a pump is lowered to the bottom of the piezometer
and the cavity is flushed, t remove puddling effects, by
puffiping out the in-seeping water (fig. 9).

In more permeable

soils the cavity is quickly flushed out while in tight clay

soils the water may seep in so slowly that it is necessary to
apply suction to the pipe to speed up the process (fig. 10).
Permeability measureaients are then aade by pumping the
water out of the piezometer and measuring the rate of rise of
the water in the piegoaeter.

An electrical probe which works

on the principle that an electric circuit is completed when

the end of the probe touches a water surface, may be used to
measure the water elevation in the piezometer.
fhe level of the water table must also be known for the
calculation of the soil permeability.

In highly permeable

-39-

10

Applylag Suetion to the Piezometer,

Upper Stopper la for a 2-incto
Piezometer.

40-

soils It takes only a few minutes for the water table to

establish itself.

In tight olay soils it may take hours ox

even days for the water table to reach an equilibrium state.

When the laeasurement is coiapleted, the piezometer is
removed from the soil by means of a soil tube jack.
2,

Field tests
Preliminary tests on prairie soils under continuous pas

ture showed that it was essential to remove the surface sod

before installing the piezometer.

Forcing the piezometer

through the sod caused a ball of roots and soil to form upon
the end of the piezometer.

This plug compressed the soil

around the end of the pipe and reduced the permeability.

Two types of ai;^ers were tried in field tests; one a
regular carpenter's bit with the sharp edges ground off, and
the other a double-twist-type soil ai^er.

In heavy soils

there was less tendency for the saturated soil to slip off
the carpenter's bit than off the soil auger.

Other than that

no particular advantage was noted for either auger.

It was

found advisable to use small-diameter pipe (3/8 in.) for the

handles of the auger so that it was impossible for the operator
to take large bites of soil.

Soil puddling and compression

were minimized by taking small auger bites.

The auger shaft

was conveniently marked with brass tips so the operator could

tell the depth of the hole.
Any size of pipe could probably be used for a piezometer

-41in these tests.

In the first field trials 3/4" iron water

pipes were used and were satisfactory for the iBeasurement, but
their weight made them difficult to transport in the field.
addition, water pipe corrodes very easily.

In

The pipe that proved

most satisfactory was thin-walled electrical conduit, which is

light-weight, strong, and resists corrosion.

Three diameters

of conduit were tried, 1.0, 1.5, 2.0 inches and all of them
worked satisfactorily.
10-foot lengths.

Conduit is available oommercially in

These lengths were cut in half and one end

, was beveled to provide a sharp cutting edge.

marked at 1 foot intervals with paint.

The pipe was

This was not very

satisfactory since the paint wore off after a little usage.

A more satisfactory method of marking the pipes would be to
cut a shallow groove around the circumference of the pipe
with a pipe cutter,
A driving head made of steel was placed on the piezom
eter while it was being driven into the soil to protect the
piezometer from the battering action of the maul (See fig. 8).
The puap used to puap the water out of the piezometer
was a small pitcher

coiimonly foxmd on farms (See fig. 9).

A pipe connection extended above the level of the pumping

cylinder so that the ptamp would not lose its prime between
operations.

In field tests a suction of about 0.75 atmos

phere was developed by the pump.

It was found necessary to

soak the leathers on the pumping piston in water before using

the pump.

No other pumps were tried and no particular advan

tage iB claimed for the pwmp here deseribed although it should

be pointed out that thie pu^ is of rugged construction and no
difficulties were encountered in its use in the field.
Plastic garden hoae with an Inside diameter of 17/32 inch^
and outside diameter of 15/16 inch worked very satisfactorily,
fhe inside diameter was large enoi^ to pass sand-and small
gravel and the wall was sufficiently rigid to stand up under
the suction.

Rubber tubing was not satisfactory because of its

small inner diameter and because of the flexibility of the walls

fwo types of electric probes were tried.

The first con

sisted of a wooden rod with an embedded wire.

One end of the

wire protruded from an end of the, rod.

The other end of the

wire was connected in series with a milliamiaeter, a 33-volt

hearing-aid battery* a limiting resistor and a brass collar
that fit ow@T the end of the piezometer.

Ihen the wire pro

truding from the end of the wooden rod touched the water sur
face in the piezometer the circuit was completed causing^ the
milliammeter to deflect.

The wooden rod was graduated in

inches and a set screw in the brass collar enabled the oper
ator to fasten the rod at any elevation.
leadings were faade in the following ma^er; the rod was
set at aorae predetermined height and the operator waited
til the rising water in the piezometer reached the rod.

un
A

stop watch was thegi started and the rod teoved up on inch
(or any desired distance).

lh.en the water level reached the

-43new elefatioa as Indicated by the deflection of the milliaffi-

meter, the first stop wateh was stopped and a second stop
watch started,

fhe reading on the first stop watch

recorded

and it wes reset to zero so that it could be used again to take

the third reading.

In this way it was possible to get a whole

series of aecurate readings using two stop watches.

The collar attachment that fit over the end of the pipe
also held a small horizontal square of plywood that served as
a platform for the ffiilliaaaaeter, the battery, and a notebook.
The rod-type of probe was not very satisfactory for two
reasons, (l) it was awkward to carry in a oar and in the field,

(3) it took an appreciable length of time to reset at a new

elevation.

Because of these liaitations a new electric probe

waa designed following, somewhat, the design of CSiristiansen

(15).

A fishing reel (Oreno So. 1165) was fastened on a

square of plywood (See fig. 11) and a graduated length of radio

test lead wire was wound on the reel.

A weight consisting of

an insulated brass sleeve fit over the end of the wire that
dropped down inside of the piezometer and the other end of the
wire was hooked onto the reel.

There was sufficient friction

in the catchment aechanism of the reel to hold the wire at any

desired elevation.

The reel-type probe proved to be more con

venient to operate in the field and was ouch more compact than
the rod-type probe.

It could be operated with one hand.

The soil tube Jack used to reruov the piezometer from the

-44-

fig. 11.

leasmriag later EleTation with

Reel-type Slectrio Protoe.

Fig. 12,

RoTiiig a 1-iach Piezometer with

Teihaeyer Soil Tube J&qTc.

-45groimd was the standaid eihmyer soil tube ^ack.

Special

grips were made to accomodate the 1 1/2 and 2-inch piezom

eters and the rollers were removed from the yoke when these
larger grips were used.

A grip made for 3/4'' water pipe was

also used to pmll the 1 inch conduit (See Jtig. 12).

Although

the pipe was too large for the grip, no difficulty was encoun
tered in its use.

In all cases the soil tube jack performed

satisfactorily, removing pipes up to 2 inches in diameter

from heavy clay soil,
3.

Laboratory procedure
The general solution for the problem of flow into a

cylindrical hole with impervious walls and with open end below
a water table, was given by lirkham (41).

The expression

obtained for the permeability may be written in the form

(fig. 7)
X - "TT

[linCcl-id.VCol-ya.)]
A (w-tj

where I s soil permeability (quantity of water discharging

in unit time through a unit length of soil of unit cross sec
tion, the ends of which are at unit difference in hydraulic
head), R inside radius of pipe. In s natural logarithm.

The forraula as presented here differs from that

originally given in that the viscosity and density of the
soil water and the acceleration of gravity are luaped into
the permeability constant. Also, the quantities in the
argiiment of the logarithm are revised to account for in
ward seepage rather than outward seepage as originally
described.

d * rJeptli of pipe below water table,,

= depth below ^ater

tftbla of TOter in p.ip@ at tijoe t. fg dspth below water

tablt of Bter in pipe at time tg., tg-tj^ time required fox
water to ri.se froia

to fg, A ss a fuiictio.a of the geometry

f the flow sfsteia iiaviag the physical dl.oeasioa of length.

The qmntity A depeade out the siae and sh8.pe (r,w) f

the eavlti" beneath the pieKoatter and also depends on the
distance from the oairitf to the wfi-ter table, d, and the distsjice to an iBipere.abl layer, s, te aay call A the A-function,
Phyaioally A 1 a sort of conductaaee.; it is the amount
of we.ter which will flo into the oavity in unit time if the
soil peraeability is unity and a mit difference of hydraulic
head exists between point# in the soil ftt the water table and
points in the soil over th wallt of the cavity.

The .A-fuac-

,tion is the only quantity in equation (8) which cannot be

readily measured or otherwise obtained and we roust resort to
the electric.analogue for its evsluatioa.

Frevert (2?) de

veloped the three-difflensional analogue which waa used in this

work and his thesis coataias a deta.iled description of the
construction and Qp.r&tioa of the analogue.

Briefly the

laboratory apparatus oo.asisted of a stock-watering tank 68

laches in diameter and 30 laches deep, the bottom of which waa
lined with 26-gauge copper sheets. (See fig. 13) the ta.nk was
filled with tap water which proved to be sufficiently conduc
tive for the teats..

The c^ptr bottoa of the tank was every-

~47

h
'ouKcice of
El(Zx:trc'tLjte
tZleciro cli

Fig, 13.

\ooo Ctjcl
Altar-nqto'

Proba
Copper Lined Tank-

Circuit for Location of Equipotential

Surfaces - Frevert (27).

"burfcice of
ELl^^ct rol Ljtt:
E l;ctrod<z
Q
Copper Bo-ftomTank- L

(2he-o'5,tat-

Mc-cid

onc-s

Variable

\ooo CLjclii^ Al+iz-rnotor

0
Fig, 14,

Circuit for Determination of A-function Frevert (27),

20

UJ

100

200
T I M E

''If* M*

300
I N

400

500

S E C O N D S

tf ffM l&tii te
Ms3r Ml&HmMp
Mmm
aai f i m t

600

wkere at the same potential and corresponded in tbe field

case to a stationary water table.

The tap water ishicii was the

electrolyte corresponded in the field to the soil which ccmdmets water.

The water-air smrfaee throi^h which electricity

would not pas had it analogy in an impermeable layer in the

soil.

A dielectric rod made of plastic corresponded to the

imperiaeable walls of the piezometer and a right cylindrical,

copper electrode attached to one end of the dielectric rod
was analogous to the aTity beneath the piezometer.

The

dimensions of the dielectric rod and the attached electrode

were in the same proportions as were the dimensions d and 2r
in the field,

fhe l-fmetion was coa^uted from the foraula

derived by Kirkham (41).

vn
^ere
r 5= radius of the soil ca-rity
S5 radiws of cylindrical electrode
cr specific condttotiTity of the tank electrolyte
-TL = electrical resistance between the cylindrical
electrode and the copper tank bottom
k non-indmctiTe Wheatstone Bridge was used for measuring
with 500 cycle AG current at 6 ^olts for the power supply.
Sarphones were used for balance detection*

-50-

4.

Results M usM mm
la the first field tests the piezometers f/ere driven into

the soil withomt any preYloms augering.

Headings were takes

interfflittently over a period of two weeks,

fhe perffleability

was fomnd to inerease ^th smcoessive flushi)^s of the cavity

dme to a loosening of the compressed soil beneath the piezom
eter.

fhe ffiaxinaffl valtte obtained was somewhat lower than* but

approximately the same as, the value obtained with the method
of successive angering.

The aorrtspondeaoe of these two

figures is interpreted to mean that the successive augering

BiethcKi of installing the piezometer minimizes soil compression
and puddling.
The spring of 1948 was relatively dry and by the time the
method of measurement had been perfected, there were few spots
of uniform soil with a high water table at which measuremeats
could be made.

Many of the readings were taken at sites that

were very variable in permeability due to natural soil varia

tions.

Site IJ

May 16, 1948

Site description: Tbe soil was a Webster silty clay loam. The
tests were soade on Black's Seed Farm about 3 miles south of Ames.
The Webster soil was formed from glacial till material and the
surface soil, which was in permaneat pasture, supported a dense
growth of grass that was matted into a heavy turf. The top 24
inches became progressively lighter in color, grading into a
grayish mixture of sand, clay and small gravel at 3 to 4 feet.
Table 2
Field Test Data

>ipe
lo.

Inside
Diameter
of Pi'oe
(ia.1

Depth
Saj^le
{feet)

A(in.)

t2-tx

K(in/day)
corrected
to 20O C

1
2
3
4

0.756
0.756
0.756
0.756

2
2
2
2

11.0
11.0
11.0
11.0

0.197
0.140
0.159
0.176

725
514
586
648

1
a
3
4

0.756
0.756
0.756
0.756

3
3
3
3

11.0
11.0
11.0
11.0

0.312
0.212
0.332
0.290

1150
784
1222
1069

1
2
3
4

0.756
0.756
0.756
0*758

4
4
4
4

11.0
11.0
11.0
11.0

0.310
0.425
0.368
0.294

1144
1569
1355
1081

SiMmary of Field Data

Depth
Sample
2
3
4

Average
Permeability
(in/dav)
618
1056
1287

Standard
Deviatim
(in/dav)
90
192
221

-52Sit 111

Juae 10 1948

The measmremeats were made on the South Shore of Little lall

Lake, Jewell, Iowa in a pasture, the soil was a Webster silty
loaia s,nd was similar to the soil at Site I.
fable 3
Field fest Data

Pipe
lo.
1

2
3
4

5
8
1

2
3
4
5

6
1

2
3
4

5
6

Iaside
Diaaaeter
of Pipe
(Ifit)

Depth
Sample
(feet)

tg-ti

K( l n /day)
corrected
to 20 C

0.0725
0.0490
0.0573
0.0790
0.0705
0.0733

788
532
622
860
767
798

A(in.)

1.5
1.5
1.5
1.5
1.5
1.5

2
2
a
2
2

15.0
15.0
15.0
15.0
15.0
15.0

1.5
1.5
1.5
1.5
1.5
1.5

3
3
3
3
3
3

15.0
15.0
15.0
15.0
15.0
15.0

0.0805
0.1045
0.0815
0.0876
0.0805
0.0745

878
1140
888
955
878
813

1.5
1.5
1.5
1.5
1.5
1.5

4
4
4
4
4
4

15.0
15.0
15.0
15.0
15.0
15.0

0.0773
0.0745
0.0750
0.0795
0.1120
0.1060

840
810
814
862
1218
1150

Summary of Field Data

Depth
Sas^le

2
3
4

Average
Permeability
(in/day)

Standard
Deviation
(ia/day)

728
925
949

114
207

116

52^
Site HI; July 10, 1948
file Bieaaureaeats were made ia a permaneat pasture located 6 miles
eaat of fhitiag, Iowa, near tbe relocated channel of the Little
Siomx River* fhe aoii resembled Luton olay at the surface tout
the subsoil was much lighter than typical Luton. The surface
was a heay plastic clay to a depth of 2o inches becoming progre88i"ly lighter in texture at greater depths. Layers of pure
sand were encountered at 4 feet.
Table 4
Field Test Bata
Inside
Diameter
of Pipe
(in.)

Depth
Sample
(feet)

A(in.)

InU-yi)
U-y2^
tg-tl

1
2
3
4
5
S

1.0
1.0
1.0
1.0
1.0
1.0

3
3
3
3
3
3

12.3
12.3
12.3
12.3
13.3
12.3

0.000326
0.000324
0.000287
0.000350
0.000350
0.000358

1
2
3

1.0
1.0
1.0

4
4
4

12.3
12.3
12.3

0.0112
0.0116
0.0112

Pipe
lo.

Summary of Field Data

Depth
Sample
(feet)

A-rerage
Permeability
(in/day)

Standard
Deviation
(in/dav)

3
4

1.82
63.6

0.147
1.30

K(in/day)
corrected
to 20 0
1.79
1.78
1.51
1.92
1.92
1.97
61.8
64.2
61.8

SM
Site Ii

Sept. 12-13, 1948

The meaaurefflents were made in a peifflanent pastiire adjacent to

an old slougliway of the Missouri river south of Oaawa, Iowa.
The surface soil was a grayish broa sticky clay. Several thin
sand layers were ncoantered in the soil and the subsoil waa
extremely variable.
fable 5
Field fest Data

Inside
Diaaeter
of Pipe
(in.)

Depth
Sample
(feet)

1
a
3

1.5
1.5
1.5

2
2
2

15.0
15.0
15.0

0.000662

1
2
3
4
5
6

1.5
1.5
1.5
1.5
1.5
1.5

3
3
3
3
3
3

15.0
15.0
15.0
15.0
15.0
15.0

0.000790
0.04190
0,000681
0.03234
0.0150

8.1
430
7.0
330
0.0
150

1
2
3
4

1.0
1.0
1.0
1.0

3
3
3
3

12.3
12.3
12.3
12.3

0.0828
0.1061
0.000434
0.001069

470
580
2.4
5.9

1
2
3
4
5

1.5
1.5
1.5
1.5
1.5

3.5
3.5
3.5
3.5
3.5

15.0
15.0
15.0
15.0
15.0

0.0001965
0.3848
0.000185
0.0255
0.000645

2.0
390
1.9
260
6.6

1
2
3
4

1.5
1.5
1.5
1.5

4
4
4
4

15.0
15.0
15.0
15.0

0.0000110
0.000320
0.0000110
0.00120

Pipe
Ho.

A(in.)

inid^dO)
yd-yaj

K(in/day)
corrected
to 20 G

0.003495

6.8
0.0
35.6

.11
3.3
.11
1.2

Table 5 contizmsd
wOCQ

III
a
3
3
3.5
4

Swm&Ty of Field Data

Average
Permeability
(in/dair)

ai.2
154.7
264.6
132.1
1.2

Standard
JDe-riation

Comment

(In/day)

20.8
175

304
182
1.5

1.5
1.5
1.0
1.5
1.5

inch
inch
inch
inch
inch

pipe
pipe
pipe
pipe
pipe

5e-

Site ;

Sept. 23, 1948

Tests were made on the north shore of Lake Ahquabi near

Indianola, Iowa, fhe soil has been classified as Wabash
slit loam which consists of 10 inches of silt loam surface
soil over a heavy dense clay subsoil. Under native condi
tions this site wai covered with a forest of hardwoods but
when the lake was constructed the trees were removed along
with about a foot of the surface. The soil remaining waa
much disturbed and contained any tree roots, v
fable 6
Field Test Data

Pipe
lo.

Inside
Diameter
of Pipe
(in.)

1
2
3
4
5

1.5
1.5
1.5
1.5
1.5

1
2
3
4
5

1.0
1.0
1.0
1.0
1.0
1.0

l.S
1.5

1
2
3
4

1.5
1.5
1.5

5
6

Depth
Saffiple
(feet)

A(ln.)

2
2
2
2
2

15.0
15.0
15.0
15.0
IS.O

0.000309
0.000415
0.000191
0.000105
0.000387

3.2
4.2
1.9
1.1
3.8

2
2
2
2

12.3
12.3
12.3
12.3
12.3
12.3

0.000431
0.00166
0.000350
0.00136

2.4
4.1
1,9
7.4
0.0
0.0

0.0000548

.56
0.0

0.000458
0.000510

4.7
5.2
0.0

'

K(ln/day)
correct^
to 20 0

tg-ti

3
15.0
3
15.0
-Hit a rock-
3
15.0
3
15*0
3
15.0

Suaaary of Field Data

Depth
Sample
(feet)
2
3
3

Average
Permeability
(in/dav)

Standard
Deviation
(in/day)

Comment

2.85
2.58
2.08

1.33
2.67
2.65

1.5 inch pipe

1.0 inch pipe
1.5 inch pipe

-57Site ?I;

Oot. 30-31, 1948

Measmxements were made in the ibottom of a roadside ditch near

Lake Oornelia, Iowa, fbe soil was of glaoial origin and con
sisted of 1 to 8 feet of peat o-rer a gray mottled clay containing
grawls and cobbles, fhe soil was very variable and pockets of
sand aad gravel were mixed in with heavier aaterial.
Table 7
Field Test Data

Pipe
lo.

Inside
Diameter
of Pipe
(in.)

Depth
Sample
(feet)

A(in.)

id-ra)
*8-*l

K(in/day)
corrected
to 20 0

1
2
3
4
5
6
7

1.0
1.0
1.0
1.0
1.0
1.0
1.0

3
3
3
3
3
3
3

12.3
12.3
12.3
12.3
12.3
12.3
12.3

0.00356
0.0520
0.0130
0.0219
0.00943
0.0680
0.0820

19.6
286
72.0
120
52
385
450

1
S
3
4

2.0
2.0
s.o
2.0

3
3
3
3

17.0
17.0
17.0
17.0

0.0163
0.00456
0.01154
0.00610

260
73
185
100

1
3
3
4
5
6
7

1.0
1.0
1.0
1.0
1.0
1.0
1.0

4
4
4
4
4
4
4

12.3
12.3
12.3
12.3
13.3
12.3
12.3

0.01048
0.00326
0.00176
0.00130
0.00326
0.00579
0.00041

1
2
3
4

2.0

3.0

17.0
Hit rock-^
-Hit rock
17.0

58
18
9.7
7.2
18
32
2.3

0.00650

104

0.00181

29

-58-

fatole 1 continued
Stwinary of Field Data
Depth
Sanplc
(fett)
3
3
4
4

A-^erag
Perffleability
(in/day)
197.8
154.5
20.7
66.5

Standard
Deviation
(in/day)
174
85
19.1
52.9

Comment
1.0
2.0
1.0
2.0

inch
inch
inch
inch

pipe
pipe
pipe
pipe

-59.

Site 11:

June 11, 1948

Measureaents were made on an area of prairie soil in Howard

County that Ma been maintained in its virgin condition by the
State Oonser'ration Gommission. The soil has been classified
as Clyde silt loam and has a shallow surface layer about 5
inches thick. The subsoil showed tremendous variation over
sffiall distances; clay pockets and sand pockets oeeurring
within a few feet of each other.
fable 8
Field Test Data

^pe
^0.
1
2

3
4
5
6
7
8
1
2
3
4

1
3
3
4

6
6
7
8
1
2.
3
4
1
S
3
4

Inside
Diaaeter
of Pipe

Depth
Sample
(feet)

A(in,)

*2-l

(in,)

1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
.756
.756
.756
.756
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
.756
.756
.756
.756
1.0
1.0
1.0
1.0

2
2
2

0,01052
0.01230
0.01340
0.00886
0.01900
Hit a rock~
0.0368
0.01058

K(in/day)
corrected
to 20 0
58
68
74
49
110

2
2
2
2

12.3
12.3
13.3
12.3
12.3
13.3
12.3
12.3

3
3
3
3

11.0
11.0
11.0
11.0

0.0146
0.00002

3
3
3
3
3
3
3
3

12.3
12.3
12.3
12.3
12.3
12.3
12.3
12.3

0.00427
0.00496
0.000886
Hit a rock
0.0128
Hit a rook
^Hit a rock

4
4
4
4

11.0
11.0
11.0
11.0

0.00000115
0.00000176
0.0000304
Hit a rock

0.004
0.006
0.11

4
4
4
4

12.3
12.3
12.3
12.3

0.000531
0.00305
0.00000697
^Hit a rock

2.9
16.8
.039

0.00854

205
58
500
0.07
0.0
294
23.5
0.0
27.3
4.9
71.5

Table 8 coatinued
SuffiBJary of Field Data
Depth
Sample
{feeti

Average
Permeability
(in/da)

Standard
Deviation
(in/day)

2
3
3
4
4

88.8
198
85.2
0.04
6.6

17.3
357
88.3
0.085
12.7

CoBunent
1.0 inch pipes
.756 inch pipes
1.0 inch pipes
.756 inch pipes
1.0 inch pipes

-61Sit VIII:

Juae 10-11, 1948

ftsts were mad# in a field aortii of the Howard County &rm

on a soil similar to that found at Sit HI. Sand and clay
-pockets were found throughout the profile.
Table 9
Field Test Data

Pipe
lo.

Inside
Diameter
of Pipe
(in.)

Depth
Sample
(feet)

A{ln.)
tg-tx

K(ln/day)
corrected
to 20 C

1
3
3
4
5
6
7

1.0
1.0
- 1.0
1.0
1.0
1.0
1.0

3
3
3
3
3
3
3

12.3
12.3
13.3
12.3
12.3
12.3
12.3

0.00280
0.02215
O.OB980
0.05430
0.04355
0.05100
0.06550

15.4
122
495
300
240
281
360

1
2
3
4
5
6
f

1.0
1.0
1.0
1.0
1.0
1.0
1.0

4
4
4
4
4
4
4

12.3
12.3
12.3
12.3
12.3
12.3
12.3

Hit rook
^mt rock
0.01886
0.02410
0.01214
0.001008
0.03910

104
133
67
5.6
216

Sawaary of Field Data

Depth
Sample
(feet)
3
4

Average
PermeaMlity
(in/day)

Standard
De"yiation
(layday)

259
105

156.4
74.5

-62-

Site IX:

Oct. 15, 1948

flie measurements were made on Marion silty clay loam which is

a terrace soil. The sit was located on a flat just below a
hillside seep near Fairfield, Iowa and the soil was quite uni
form.
Table 10
Field Test Data
K{in/day)
corrected
to 20 C

Inside
Diameter
of Pipe
ipi
-Mm 11

Depth
Sample
(feet)

A(in,)

1
2
3
,4

3.0

2
2
2
2

17.0
17.0
17.0
17.0

0.00996
0.00955
0.00641
0.00595

159
153
103
95

2.0
2.0
2.0
2.0

3
3
3
3

17.0
17.0
17.0
17.0

0.0145
0.00643
0.00716
0.00488

232
103
114
78

1.0
1.0
1.0
1.0

3
3
3
3

12.3
12.3
12.3
12.3

Pipe
lo.

2.0
3.0
2.0

3
3
4
1
2
3
4

ln4Mll
(d-yg)
t3~ti

0.0243
134
0.0289
159
0.0145
80
Damaged pipe

SuHifflary of Field Data

Depth
Sample
(feet)
2
3
3

Average
Permeability
(in/day)
127
132
124

Standard
De'S'latioii
(In/day)
33
68
40

Comment
2.0 inch pipe
2.0 inch pipe
1.0 inch pipe

-63Site 1:

April 1-2, 1949

Measurements were aade on the Davis Gounty KxperiBaental Farm

near Bioomfield, Iowa, fhe soil VB.S Edlna ailt loam and con
sisted of a 10 inab surface silt loam underlain by a heavy,
dense clay subsoil. There was a 1 inoh leached glel layer at
10 Inches.
fable 11
Field feat Data

>ipe
lo.

Inside
DiMeter
of Pipe

Depth
Sai^le
(feet)

A(ln.)

(in A

(d-ya)
tg-tl

K(in/day)
corrected
to 20 C

1
2
3

1.0
1.0
1.0

1
1
1

12.3
12.3
12.3

0.00727
0.00431
0.00251

53.1
34.5
20.2

1
2
3
4

3.0
2.0
2.0
2.0

2
2
2
2

17.0
17.0
17.0
17.0

0.00100
0.00024
0.00040
0.00018

27.0
5.7
9.2
4.2

1
3
4
5
6
7
8

1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0

2
2
2
2
2
2
2
2

12.3
12.3
12.3
12.3
12.3
12.3
12.3
12.3

0.00055
0.00170
0.00157
0.00114
0.00093
0.00061
0.00125
0.00165

44.5
14.0
12.7
9.2
7.6
6.6
10.1
13.3

1
3
3
4

2.0
2.0
2.0
2.0

3
3
3
3

17.0
17.0
17.0
17.0

0.000448
0.000075
0.0000064
0.0000257

10.4
1.8
0.15
6.0

1
2
3
4

1.0
1.0
1.0
1.0

3
3
3
3

12.3
12.3
12.3
12.3

0.000001
0.000000
0.000030
0.0000003

0.010
0.00
0.24
0.0024

1
2
3

3.0
2.0
3.0

4
4
4

17.0
17.0
17.0

0.000029
0.0
0.0000013

0.67
0.00
.030

-64.

Table 11 ooatlamed
Suamary of Field Data
Depth
Sample
^feet)
1
2
3
3
3
4

Aterage
Fexaeabllity
(layday)
35.9
11.5
14.7
4.S
0.063
0.23

Standard
Deviation
(ia/day)
11.2
9.5
4.1
4.6
0.205
0.535

Oommeat
1.0
2.0
1.0
2.0
1.0
2.0

inch
inch
inch
inch
inch
inch

pipe
pipe
pipe
pipe
pipe
pipe

-65Tke permeability figures at Sites I and II are very high

and according to Muskat (51) correspond to the permeability of
laedium sand.

On the other hand Hissink (36) and Hiasink and

Hooghoudt (3?) in laboratory tests made on undisturbed soil

samples have obtained peraeability figures of from 100 to 3000
inches per day on soils that contained about 65^ clay.
The standard deviations of the readings at Sites I and II
lie between 10 and 20^ of the average.

Although a standard

deviation of 10 to 20^ may be considered large in many physical

measurements, it should be remembered that measurements of soil
permeability, taken by coamonly used methods, often vary by
taore than 3000^ (14).
The measurements at Site III rejroduced themselves very
well with a standard deviation of less than 10^ of the average.
However, at Site IV, soil variations caused such a wide range
in permeability values that no conclusions can be drawn from
the results.
At Site an attempt was made to check the results obtained
by using piezometers of two different diaaieters (1.0 and 1.5
inches).

Althoiigh the standard deviations are quite large, the

averages, 2.58, and 2.08, inches per day, are close.

The vari

ations in the permeability are attributed to the soil hetero

geneity.
Sites ?, VI, VII, VIII all had very variable soils condi
tions and these variations are reflected in the permeabilities

-66obtalned.
The soil at Site IX was quite uaiform and offered an
opportunity to check the validity of the method by comparing
the results obtained with a 1 inch piezometer with the results
of a 3 inch piezometer.

The 1 inch piezometer at a depth of 3

feet beneath the soil surface gave an average value of 124 t 40

inches per day while the 3-inch piezometers at the same depth
gave an average value of 133 68 inches per day which is con
sidered a satisfactory check.
Measurements were made on the Edina silt loam to see
whether or not the method would work on soils as heavy as the
subsoil of the Edina.

The results in general were not satis

factory and the method will have to be modified for use on

heavy soils.

Water moves so slowly through these heavy soils

that it is difficult to flush out the cavity with in-seeping

water.

One possible method of flushing out the cavity would

be to lower a hose into the cavity and direct a slow stream of

water into the cavity.
of the piezometer.

The water would then flow out the top

Reeve (54) has tried this method with soils

in the West.
Leakage along the walls of the piezometer may be serious
in some soils and when it occurs, the values obtained for the
permeability will be too high.

Since most soils vary in color

with depth, any leakage could be detected by examination of the

color of the suspended matter pumped out of the cavity.

Ho

leakage was encountered in any of the tests reported here.

-67-

As additional evidence to support tbis conclusion, it was

noticed that ^ea the piezometers were pulled out of the
ground, they had a tight compressed layer of soil adhering to
their surface,

fhis compressed layer would prevent leakage

along the wall.

It is difficult to prove the validity of the measurements
obtained by a field raethod such as the one described here.

By

perforraing the operations of installing the piezometer in ex

actly the same way each tiae and by making all of the readings
in the same way, it is possible for the operator to reproduce
his results satisfaetorily and still not have the true value
of the perffleability.

For instance, if the operation of in

stalling the piezoraeter were performed so that the soil be

neath each piezometer was coi^ressed about the same araount,
the readings on different piezometers might check but all of
the readings would be reduced about the same amoimt below the
true value because of the soil compression.
lo method has yet been devised to drive pipes into the
soil without producing some change in the soil itself.

Hoi^

ever, it is felt that the procedure outlined in this thesis

will iiiniaiize such changes.

As an additional safeguard, the

readings on the individual pipes were repeated until the cal

culated permeability reached a constant value.

In most cases

the first readings were somewhat lower than subsequent readings

because all of the puddling effects had not yet been eliminated.
Continued flushing of the holes eventually produced stationary

-68-

Talmes.
On the Tery permeable soils at Sites I and II, it was
difficult with the equipment used to obtain readings at a
depth of more than a few inches beneath the water table.

?ilhen

readings are taken close to the water table, any small error
in the water table aeasureiaent or any drawdown effects will
cause a very large error in the permeability figure obtained,
for these reasons it was thought advisable to take all readings
at least six inches below the water table and deeper if pos
sible.
5.

Laboratory results
Fig. 16 shows the variation of the A-function with diameter

of cavity for a 4-inoh long cavity.

This figure may be con

sidered valid only if d w, and s wj but is approximately

valid for other values of d and s as indicated below.

The data

presented in fig. 16 may be used to calculate the A-f\mction of

o&vities having other dimensions, since the A-function is pro
portional to the dimensions of the system^.
the following example;

This is shown by

The A-function for a cavity 6 inches

long and 2 inches in diaaietex is desired.

A cavity 4 inches

long to be in the saae proportion must have a diameter of 1.33

inches.

The value of the A-function for a cavity 4 inches long

%irkhaMi (41, p. 62) has proven that doubling the dimen

sions of a flow system will double the A-function, but we have
shown (fig. 18) that when d and s are large compared to w, the
value of the A-fmction is independent of d and s. Therefore,
with d and a large, doubling the dimensions of the cavity
doubles the value of the A-function, regardless of d and s.

-6930
zm

CO
LlJ

TY ^ MNC:HES LorsIG

o 20

5 10
I

<

DIAMETER OF CAVITY (INCHES)

Fig. 16,

Vaxiation of A-fuactioa with Diameter

of Gatity for a 4-inch Iong Cavity (For
Values f B
d See fext),

30

c/)
0
1

20

WIT Y 1- INCH DIA

J L

UJ
X

ER
/

3
ij.
'k
0

Fig. 17.

LENGTH OF CAVITY (INCHES)

Variation of A-fiinetion with I>e]&gtli
f Cavity for a 1-iaeh Diameter Cavity.
(For Values of s and d See Text).

-70and 1,33 Inciies in diaaeter is read from fig. 16 to be 14.0.

The value then for the oavity 6 inches long and 2 inches in
diameter is (6/4) x 14.0 inches s 21.0 inches,
fig. 17 shows the Tariation of the A-function with length
of cavity for a cavity 1 inch in diameter.

Fig. 17 may be used

like fig. 16 to compute A-values for cavities of other dimensions

than those given on the graph.

Thus, the A-function for a cavity

2 inches in diameter and 4 inches long is obtained as follows:

For a cavity 1 inch in diameter to be in the same proportion,
it vmst have a length of 3 inches and the A-function reading
from fig. 17 would be 8.5 inches.

Since the cavity 2 Inches

in diameter and 4 inches long has dimensions which are twice

those of the 1x2 inch cavity, the A-function woiild be 2x8.5 17.0, in agreement with the same value read from fig. 16.
Fig. 18 show how the A-fimction varies with d(or with s)

when w 55 4 inches, 2r 1 inch, and s 4-w + d a 24.5 Inches.

- ^ t;

This curve brings out the important fact that if d and s are
both large corapared to w, the
with d and s.

A-function valies

inappreciably

Examination of this curve shows that the value

- >

of the A-function when obtained from fig. 16 or 17 will be

correct to better than about

if d 5^ w and s

l/2w; and

that the values will be correct to better than about 4^ if

d ^Sw, s ? 2w.

If s s 0 and d w, the value of the A-f\inctioa /

will be about 25^ too large when obtained from fig. 16 or 17.
The curve is relatively flat between d w and s 1/2w.

It

^
^ ^

fL^n

A^fjA*'

I
2. Y

[212

I^

z
=8
z
o

s^w^d = 24.5 INCHES

w = 4 INCHES
2r = I INCH

g4
3
Lid>

< 0

I
8

,1 I
12

+ -! s S4 f

s+w

20

16

y:

t ^^'s

.-ik

u.j

Sf-'-' fii --

s =0

-t

'(

5 ^ ' - '

=0
24

,r < i

28

S-'

DISTANCE d BELOW WATER TABLE (INCHES)

'--T X<i, r'

fJ
Fig. 18. l^sriatloB of A-ftmction with a, for
a OaTity 4-iiiehes long and 1-inch in
DiaBieter.

PIEZOMETER

XABli

WATER

100

KX)

985
L50
ko
-70

IMPERVIOUS

LAYER

, , ,,

i//////////////n///////n un////////I N////////in////!//////m I!/>'

Fig. 19.

l<|aiptatlal Suxfaoes for a

Pitzoaeter Ga^ity.

-^'.r -f-'"'-! - . . M

-.T2ia from this portion of th cur-r that values were obtained for
fig. 16 and 17.
Strictly speakiag, if the soil consists of stratified layers,
the A-fmction depends mpon the soil permeability ia each of the
'i

layers, and mpon the dimension of the layers as well as the diffieasioa of the oa^rity.

However, if a layer is not closer to

the oavlty than the length of the cavity the effect of these
layers, regardless of their permeability will be small.

This

conclusion was verified by sifflmlating the following two extreme

field eases in the model tank.

(A)

fith d 1 foot, w a 4

inches, 2r 1 inch and s = l.S feet, an impermeable layer 1

inch thick and 4 feet in diameter was placed in the 'soil* 4
inches above, and coaxially with the cavity,

fhe A-function

was measured and found to be reduced 11.Si from the value with
out the impermeable layer,

fhe diameter of the disk was now

increased to 8 feet and the A-function measurement made again,

its value remaining sensibly the sajne.

(B) With d, Sr, and s

as in (A), a layer of "coarse gravel" 9 square feet in area

and 4 laches thick was placed in the "soil*, with the lower
surface of the "gravel* 4 inches above the top of the cavity.
The A-function was measured and found to have increased
with respect to the value with no gravel.

In these experiments,

the impermeable layer was sibilated by a disk of shellacked

plywood, and the gravel", by copper sheets bolted together
by conducting screws.

A 1/2 inch electrode sisaulated the 1

inoto piezometer.
Another ctoeck on the proposition that only the soil in a
small region about the canity contributes to the soil permeabilitf was obtained by determining (with the model) the equipotential surfaces about the "cavity" as shown in fig. 19.

The

Talue 100 is assigned as the potential (hydraulic head) at the

surface of the soil and the value isero at the cavity.

The

figure shows that 80^ of the hydraulic head is dissipated in

the soil within a distance of 1 cavity length from the cavity
and about 90^ within two cavity lengths.

Soil outside the 80

and 90 rings thus contributes little to the flow and hence to

the A-function and to the value of K.
The effect of roeke and channels in the soil on the
periaeability measureaent was investigated with the model.
Pieces of dielectric were used to simulate rocks, copper wires
to simulate worm holes, and copper strips to simulate cracks
in the soil.

If the "cracks* and *worm holes* were not contin

uous from the cavity to the soil surface, their effect was
generally small.
the A-function.

Rocks near the cavity had little effect on

These findings are essentially the same as

those reported by Frevert (27).

B,

Application of Permeability Data to a Drainage Probla

The various methods of obtaining solutions for the drainage

problem have been covered briefly in the Review of Literature,

~f4-

Of all tlie methods meiationed, the ntimerical analysis procedure

Is the one oapatole of application to the widest variety of
problem with a miaifflim of skill and labor.

Several examples

will be wor&ed out to show how the method can be applied to

drainage problems involving soils with layers of two different
permeabilities.

The equipotentlals will be determined by nu

merical methods and the amount of water flowing into a drain

tube will be determined graphically using the equipotential
plot.
1.

Procedure
Following a suggestion of Dr. R. . Gaskell, Department

f Mathematics, Iowa State College, the liiebmaan procedure was

used in the analysis.

Although other nsethods devised by

Shortley, Southwell and others are somewhat more rapid, the

simple iiiebfliann procedure has the advantage of being easy to
understand and simple to apply.

Once the problem has been set

up, the numerical analysis can be readily done by any untrained

person; a stenographer or student help.
mathematics is necessary.

Ho special training in

In spite of its simplicity, the

niiaerical method is powerful enoxigh to treat some of the most

coaplex problems, any of which are intractable to methods of
mathematical analysis.
As an elementary example of the Liebmann procedure, let ua
suppose that in a region bounded by the dotted lines in fig. 20
a function <j> satisfies the finite difference approximation of

fill.

loo'f^

300^

I
K 1
600 ,

400

900
8 75

866
860

1500^

1000
994
989
976

900
873
865
853

300

600

700
669
650
6 32

680

700
672
6 35

652

620

500
4 I7
389
374

500
408
366
350

500
395
3 75
367

too

700
7I6

'lOOO^

-OH

o
too G

f ummrn

40 0

-76Lapla.c*s equation and is therefore an approximate solution

of the differential equation.

may be a potential function

or it may be a stream function.

It is assumed that the values

of

are known at the boundary points A, B, 0, ..., K, L.

fhese knoTO values art indicated in fig. 30 as slant numbers.

Points a, b, c ..., h, i are interior points at which
unknown,

is

fhe initial values assigned to the interior points

are estimates or guesses of what the operator thinks the value

should be.

The closer the estimate is to the true value, the

more rapid will be the convergence to the final answer so that

any preliminary calculations or work will be well repaid in
labor saved in the final calculations.

However, any finite

assigned values whatever will lead to the correct result.

With the values knoiwx at the bo\indary points and with
arbitrary assigned values at the interior points, the problem
is now ready for solution.

The net of interior points is tra

versed moving from point to point replacing each value at the

interior points by the average of its four neighbors.

On com

pletion of the first traverse, the net is again traversed

second set of values resulting.

with

This procedure is repeated

until the values repeat themselves to the desired number of

significant figures.
satisfied.

The difference equation will then be

For example, in fig. 20 a net of nine interior

points and twelve boundary points is drawn.

The top figure

-77-,

In the col\OTi of figures at polats a, b, c,

h, 1, are

the estimates made at the start of the calculation.

The sec

ond figure in the coliirai represents the averages obtained on

the first traverse and the third figiiire, the third traverse,
etc.

fhe traverses were all made from a to i alphabetically,

thus in the first traverse of the net

0a = 1/4(1000 + 1000 -H 700 + 800) = 875
1/4(1500 + 900 -H ?00 -f- 875) 994

(p'a 1/4(1000 + 800 + 700 + 994) a 873

Improved values are always used in the calculation.

For

example, the fourth value In 0 b Is the improved value for 0' a

and in $' e the iijjroved 0' b is used.
The operator need not worry about mistakes made in arith
metic since sueceasive traverses will eliminate the errors.

The

net is traversed until stationary values result at the interior

points, thus providing a solution of the Laplace equation within
the given region to three significant figures.
a.

Drainagge of aoil with uniform permeability

To show how the Liebmsnn aethod could be g^plied to a

drainage problem, a sisple case for \fhlch the analytical solu
tion is known was solved and the result obtained by the LleblaaJMi procedure was compared to the analytical answer.

The flow

into a tile drain embedded in a soil of uniform permeability

was studied.

At first a solution was attempted for a single

drain tube in an infinite medlura.

That Is, the soli was assumed

-78-.

to extend dowawards and sideways for an infinite distance.

In

solving the problem it was necessary to determine values of the

function at an arbitrary boundary drawn to confine the problem.
The values on the boundary were determined by extrapolation
from the three points adjacent on a line extending into the
interior of the region.

The

rate of convergence to the true

value was so slow as to render the method impractical and it

was decided to bound the lower side of the region by an im
permeable layer and to place squally spaced tile drains on
either side of the tile considered,

fhia confined the region

with boundaries on which the value of the function or the value

of its norma.1 derivative was known.

Due to symmetry, it was only necessary to consider a

region extending from one tile drain to a plane intermediate
between the two tile drains.
It was assumed that the spacing of the tiles was uniform
and regular.
function,

At the surface of the water table the potential

, was every^iere given a value of lOCX).

The

po

tential function was assigned a value of zero over the sur

face of the drain tube (which was assumed to be flowing full
with no bacfe pressure and with completely permeable walls).
At points along a line extending vertically through the cen
ter of the drain, and along a line extending vertically through
a point midway between the drains, and along the surface

of

i^ermeable layer, the aoraial derivative of the potential is

the

792 ro.

That this is so can toe readily shown by a consideration

of the streamline pattern.

The line extending vertically

through the center of the drain is a streamline; so is the

line midway between the drains;and so is the surface of the
iJiperiBeatole layer,

fhe derivatiire of the potential normal

to a streamline Is zero since the equipotentlal lines and

the streamlines are orthogonal.
The solution as obtained by the Liebmaan procedure is
shown on fig. 31 by the vertical numbers at the net intersec
tions.

Values at points on boiindaries on which the normal

derivative was zero were computed by talcing one-fourth of

the sum obtained by adding the values of the two points on
the boundary to twice the value of the Interior point.

This,

in effect, reflected the Interior point to a point outside of

the boundary.
Fox (25) has developed suitable equations based on
linear interpolation for handling points near a curved bound
ary.

Referring to fig. 22.

X ^

(i>H f <^a/h
3 + </h

(10)

A comparison of the nuaierioal solution was made with the ana

lytical solution presented toy Klrkham (43) v,'hlch Is
(t>-<^yin {[posh(x-ma)/2h-cosTry/2h3 / [coshTr(x-ma)/2htcoSTy/2hlj
x-ma)/2h+co8ir(y-2d)/2h3 / [eoshTT(x-ffla)/2h-co8Tr(y-

(11)

TABLE

WATER
loa

000

846

836

1000

1000

1000

877

1000

1000

1000

1000

1000

1000

1000

1000

1000

973

96!

935

1000

849

872

896

916

932

945

954

961

966

970

972

974

975

975

687

744

796

837

870

894

912

925

934

940

945

948

950

950

507

622

708

770

815

8 49

874

892

904

912

919

924

926

927

534

646

720

772

812

841

861

877

887

8 95

901

905

906

534

624

692

744

785

813

835

852

865

8 75

882

886

887

6 27

682

728

766

792

815

832

846

857

865

8 70

870

670

648
412

383
U

560

sJ 336
n

864

894

316

297

415

4$4

460

500

560

544

564

724

608

80!

839

561

598

641

683

721

7 55

779

801

81 8

832

842

851

856

8 58

604

627

657

690

720

74 9

771

790

807

821

831

840

845

845

634

649

671

6 96

721

746

765

783

799

813

822

831

834

835

605

595
632

629
649

650

719

664

en

777

6 5^

665

682

702

722

744

762

7 79

794

806

814

823

827

829

665

674

688

706

724

742

7 60

777

792

804

813

820

825

827

669

677

692

707

724

742

760

791

803

81 3

819

823

824

///A

MP

660

661
663

665

Fig. 21.

7 76

Squipotential Plots for Case of Drainage of Uniform

Soil. See Text for Details.

-81-

li
a

' 0a

0.

-a- ^

01

. aa. fdlmt M9kW a nrvtd Bemndftyy.

-82-

wbere thie origin of the

coordinate system is taken at the

center of a tile drain and

a s distance between tile in feet
h ~ distance from water table surface to an impermeable
layer
d = depth of drain tube.
and q are determined by the boundary condition.

In this case

s 1000 and q 140.6.

The true values were obtained for several points and they
are shown m the slant numerals on fig, 21.

The close agreement

between the true values and the numerically-^obtained values is

readily apparent.

The greatest error occurs at points lying

close to the drain.

points.

The function is changing rapidly at these

It would have been possible to get more accurate nu

merical results about the drain by reducing the net interval.

At points in the interior of the region the discrepancy is less
than 1^ whereas it goes up to

at points iaiiBediately adjacent

to the drain.
Ihile the equipotential plots are of interest, of far
greater value is the fact that the flow into the drain can be
calculated from the equipotential curves,

the flow across an

equipotential line over an area A is given by the expression

<l = K f ^ A
where I the permeability
change of the potential function in a direction
aormal to an equipotential line
k the area between the points considered

(IS)

-83-

fli surface of the water table lies on an equlpotential

line and it was coa'^'eaient to cooprnte the flow across this
surface since the flow across tiia stirface will also equal the
flow into the drain tube,

fhe; potential function at the water

table was everywhere 1000 while the potential functions at a

distance 1 foot below the water table were computed and are
indicated on fig. 21.
of the soil

At each foot inter'^al along the surface

was computed by subtracting the potential

function at the 1 foot depth from 1000; multiplying the result

by 3.75/1000, since 3.75 feet is the total potential drop from
the water table surface to the wall of the tile drain; and by
dividing the result by 1 foot, which is the distance
tween the two points.

be

Since the interval between the points

along the water table was 1 foot and the thickness of the soil
considered, in the iircctioa perpendicular to the plane of the
figure, is also taken as 1 foot, each value of
plied by an area 1 square foot,
up all of the K

was multi

faking K 1 ft/day and adding

along the surface of the water table from

a point over the tile to a point midway between the tiles, a

value was obtained for 1/3 of the total flow into the drain
tube,

fhis result was then compared with the analytical result

obtained by Ilrkham*s (4 3) equation

Q 2-rrI

( t +

d - r)/ln [^sinh(TT (2d-r)/aJ/sinhifr/a-

oo

^-1)*^ In Lsinh%TTnh/a-sinh^7rr/a])/
(sinh^g-rmh/ a-sinh^TT( Sd-r)/a]

-84:-"
where I = permeability in feet/'day
t thioknea of ponded water (taken as zero in this case)
d w distance from surface of water table to center of tile
drain, taken as 4 feet
h - distance from the water table surface to the impermeable
layer, here aseumed to be 12 feet
a = distance between the tile drains, taken to be 28 feet

the result obtained by the graphical method was 6.840 cu.

ft/day while the analytical solution gave an answer of 6.644 cu.
ft/day.

The error in the numerical answer was thus 3.6^.

There were 179 points at which the value of the potential

was determined by the Liebmaan procedure and it took about
twelve hours to get the answer.

alues of the potential were

calculated by means of i:irkha*s ecpiation at 26 points in about

the same length of time.

However, (if a potential plot Is not

wanted) the answer for Q can be obtained in a few minutes from

Kirkham*s result.
b.

Drainage of soil with aon-'-uniforBi permeability

Solutions of problems involving soils with layers of

various permeabilities are difficult to obtain analytically,
especially if the interface between the two layers is irregular
in configuration*

The real power of the numerical method is

realized in the solution of these more complex problems.

In Southern Iowa and Missouri, there are large areas of
soils that are difficult to drain because of a dense heavy
clay subsoil.

The numerical method was used to study this

-85-

problem.

The depth to the elay layer varies in these soils

and for this probleia the depth was taken to be 24 inches, which
is probably deeper than the average.

It was assumed that the

trenching machine dmg a trench 18 inches wide and 48 inches deep.

A 6-inch tile was laid on the bottom of the trench half embedded
in the subsoil.

The material used to fill the trench was assumed

to have the saiae permeability as the surface soil and the depth
from the soil surface to an impermeable layer was set at 10.5
feet, and the tile spacing was set at 20 feet.

The problem was

solved for three different ratios of permeabilities of the sub

soil to the surface soil, t to 5, 1 to 10 and 1 to 100.
Iiebffiaan procedure wa used as before.

The

Points in the interior

of the region of uniform permeability presented no problem

but special formulas were required to deal vith points whic^
lay on the interface and points which lay close to a curved
boundary such as the surface of the tile drain.

(The net was

so arranged that the interface lay along a line of points).

Shaw and Southwell (61) have derived a formula for points on a
plane interface using a mechanical analogy to a tensioned net.
The result they obtained (ref. fig. 33) may be written
d) = (KH- KTlX'f 3 + 1>4H3(Ktt(P
To4(Ki+iil)

0a)-3a(TT'-tT) (13)
'

Using somewhat more rigorous methods, a slightly different

equation can be derived.

The derivation is based on the finite

approximtion to the solution of the Laplace differential eq\iatioa and is as follows:

The first derivatives in the x-y

-86-

1
!(S.
1
1
1

1^3
1
1

Fig. 23.

Point on an Interface

dlreotions are built up on the iaiowledg that the average

derivative betweea two poiats is approximately e<p.al to the
differeace in the value of the function at these two points
divided by the distaaee or in fig, 23.

d<l)

<t>2.-0O

4>,~<i>o

a.

<t>o~ d)^

<t)o-

' ft.

The Laplace differential equation which must be satisfied is

Kx
^ dx
between points

+ 4 K-^= 0
^

<1^2, and

permeabilities Ij

and

and K|j or
%- %

Ky between

(|>, and

it equals Eij.

is the average of the

^11

is equal to K| while between (;6o

ad

Multiplying the first derivatives by the permea

bilities gives
^ K l f K a j ^ da.- (to ^ ^Kl+Kg ^ ^ <tb- <l>^ ^

Ku-^^=

Ki

CL

Kn

O-

The second derivative between two points is approximately the

difference in the first derivatives at these points divided by
the distanoe between them.

"8S

ax

j_
(^)
(^)- ir (^)
a. V 2. / ^ ^

'^LCKI + Kit) (4>z.+ <J^H) - 2.(KX + Kn)^6o]

^ (K,

- i l^i 4^1 + Kn ^3 - CKH +

iAm

<^oJ

Substituting in (13) ma. taking a = 1

(J) - (Ki-f Knl ((f>2.-t- <^m) 4 S. CKt fi "F Kg <^)
{14)
H (Kir +Ki)
A review of tlie literature revsaletil no derivation for a
corner poiat sucij as oceutt at tlie upper and lower edges of tbe
treneli out in the subsoil.

Referring to fig. 24 and 25 the

derivatiires are built mp as before and the solution for the

upper point (fig. 23) is
(b - ^ Ki f
<i^w) 4- (Ki > Kg)((t>t 4 (l>i)
"
GKi -faKa
The equation for the lower point (fig. 34) is

(15)

0 . 2.Hti ((tjL'h^) -I-(K^-^ KrT)^(fe-f-(^)

(16)
4Ktt+aKl
With the aid of these special formulas for points on the
interfaee and near a ourved boundary, the problem was solved
b|r the Liebfflann prooedure.

At first a very coarse net was

%
00

<p/l

0,

i4.

(f)A

SIBI

at lipit

S4it

0o
Hi
I0z

fig# is#

ftt at %^mw 0@iaap,

ft*i.

-solaid dowa over the region and solved.

Then finer nets were made

toy smtodlviding the net intervals of the coarser net.

The values

obtained as solution of the ooarse net were used as starting

values for the finer nets.

Fig. 26 illuatrates the use of four

different net intervals in solving one proMcm.

At points close

to the drain tube the fmotion changed rapidly and a fine net
was required to give the desired accuracy.

At points away

from the drain tube, the desired aocuraoy was attained by Uie
coarser nets,

the rate of convergence was speeded up materially

by startii]^ with the coarsest net possible.

Ihen a finer net is

constructed by adding one point at the center of each four

points, the resultant net is square but is diagonal to the first
net.

Since lapla*a equation is invariant on rotation of the

axes, the saw averaging formula applies to values on the sew

net.
It toofc about 24 hours to solve the problem of permeabili
ties in the ratios f i:i with the aid of a Monroe calculating
machine.

The solution for the .'1 problem was used as a starting

value for the 10:1 problem, which was solved in 16 hours.

The

100I1 problem took 11 hours to solve.

Althoxjgh no analytical solution is available to check the
accuracy of the numerical solution, a rough check was obtained
by measuring the normal tangents of the equipotential lines
with respect to the interface,

the ratio of the tangents in

the two isedia equals the ratio of the permeabilities (See

Bewley 6, p. 68).

1000

WATE

TABLE
1000

1000

1000

37 942
5M Ml

SI"""'264 296 399 |A81 ^ 60l^ 5C 990 7Y 769

3L'.*.ii.icamT3[:cnii:*jr^5rrmr.tT.iH

^^ffT ^57? SiU fiSi 686 1

0\2O

427 449 479 i

544 575 606 636

B60 681

///////A M
Fig. 36.

OUS////////LAYER"

Equipotential Plot for Drainage of a Two-layered

Soil.

-92-

WATER

TABLE

Z/////// IMPERVlbUS^/-^////LAYER'/////Zy/^///

Fig. 27.

Equipotential PIot for Drainage of a Two-layered

Soil.

-93"

WATER

Fig. 28.

TABLE

Equipotential Curves for Drainage of a Two-layered

Soil.

-94Table 12
Results of lumerical Analysis
%/%l

Q( ft3/day)

!
5
10
100

6.289
2.905
3.667
2.586

100.0^
46.1
42.4
41.1

Calculated
It is clear from the results that when the ratio of the
permeabilities Ej/Kji is 8 or greater, very little water flows
through the subsoil,

ihen the ratio of the permeabilities is

100, the subsoil is for all practical purposes ia^serraeable siace

the equipoteatials ia the Ej raedim strike the interface at
slightly less thaa 90.

If it were completely iiiQ>ermeable, the

angle would be 90 .
We may conclude, therefore, that on the Edina soil, which
has a peraeability ratio of at least 10, very little flow takes
place thxoxigh the subsoil when tile drains are placed at four
foot depths.

Most of the water must flow laterally through the

surface soil until it reaches the more permeable trench-fill

material.
Kirkham (42) has solired the problem of seepage into a
drain tile half eiabedded in an iaiperiBeable layer and his result
is
Q= J_
^

(17)
In cot irr/qh

where the symbols refer to the same dimensions as in eq. (14).

-95-^
To compare the relative advantage of placing the tile at
the bottom of a four foot trench, as was done in the problems
here solved, or of placing the tile at a 2 foot depth vdth half
of the tile embedded in the impermeable layer, the drain flux
was calculated for h d 2 feet, r .25 feet, K 1 foot/day,
t 0.

The result is
Q a 2,371 cu. ft. per day

Gomparing this result with the numericaHy-obtained answer for

the two-layered soil with permeability ratio Kj/Kn 100 (see
Table 12) it will be noted that the flow into the tile is reduced
11.1^ by placiag the tile at a two foot depth on the surface
of an impermeable layer,

if the tile \fere placed a small

distance above the imperaeable layer the flow would actually

be increased above this amount as pointed out by Kirkham (42).
The small increase in flow obtained by placing the tile at a
four foot depth would probably not justify the additional cost
of digging the deeper trench.

-96.

IV.

00NGLU8I0I AND SUMMARY

A field method of measuring the soil permeability beneath

a water table has been developed.

Piezometers are driven into

the soil below a water table, the soil is augered out of the
pipe and the rate of rise of the water in the pipe is measured,
fhe permeability is then calculated from the rate of rise by
means of an appropriate formula.
occur in the permeability formula.

Certain geometric constants

These constants have been

determined in the laboratory with a three dimensional electric

analogue of the groundwater flow problem.

The electric analogue

is used to study the effects of soil factors such as stratifi

cation, worm holes, and rocks in the permeability determination.
Field trials on several Iowa soils have shown that the
method is relatively simple, quick and accurate.

Readings can

be taken to any depth below the soil surface.

The application of permeability data to drainage has been
demonstrated using a numerical method of analysis.

Special

forariilaa have been derived to take care of points at the inter

face of two layers, each having a different permeability.

The

results obtained by the numerical methods are accurate and the

method of analysis is simple to apply.

-97-

V.

LIfERATURE ClfED

(1

Allison, L. S. Effect of ffiicroorganisms on permeability

of soil mder prolonged submergence. Soil Sci.
63:439-450. 1947.

(3

AronoTioi, V. S. and Donnan, W. M. Soil permeability

as a criterion for drainage design. Amer. Geophys.
Union frans. 37:95-102. 1946.

(3

Baver, L. D. Soil permeability in relation to non-capillary

porosity. Soil Sol. Soc. Amer. Proc. 1938/3:52-56.
1939.

(4

Soil physios.

John Wiley and Sons, New York.

1940.
(5

Bendixen, T. ff. and Slater, 0. S. Effect of the time of

drainage on the measurement of soil pore space and its
relation to permeability. Soil Sci. Soc. Amer. Proc.
1946/11:35-42. 1947.

(6

Bewley, L. V.
neering.

(7

Bodman, G. B. The variability of the permeability "con

stant" at low hydraulic gradients during saturated
flow in soils. Soil Sci. Soc. Amer. Proc. 1937/
2:45-54. 1938.

(8

Brahtz, J. H. A. Pressures due to percolating water and

their influence upon stresses in hydraulic structures.
Second Congress on Large Dams, lash. D.O. 5;43-71.
1936.

(9

Casagrande, A. Seepage throxigh dams. Jour. New Eng.

later lorks Assoc. 51:131-172. 1937.

(10

Two-dimensional fields in electrical engiMaomillan, Mew York. 1948.

Childs, E. 0. The water table equipotentials and stream

lines in drained lajad. Soil Sci. 56:317-330. 1943.

(11

T h e water table, equipotentials, and stream

lines in drained land: II. Soil Sci. 59:313-328. 1945.

(13

The water table, equipotentials, and stream

lines in drained land: IV Drainage of foreign water.
Soil Sci. 62:183-192. 1946.

-98fhe water taMe, equipotentials and streamlilies ia draiaed laad; V. The moviag water table.
Soil Scl. 63:361-376. 1947.
Ghristlansen, J. E. Effect of entrapped air upon the
permeability of soil. Soil Soi. 58:355-365. 1944.
Sojie periaeability characteristics of saline
and alkali soils. Agr. Ingr. 28:147-153. 1947.

F i r e m a n , M., and Allison, L. E. Displacement

of soil-air by Wg for permeability teste. Soil Sci.
61:355-360. 1945.

Diserens, E. trans. Sixth Gomffl.

B:188-93. 1932.

Int. Soc. Soil Sci.

Beitrag gur Bestiffliming der Durchlassigkeit

des Bodens in laturlicher Bodenlagerung. Schweizerische
I,andw. Moaathefte, 13:188-198; 204-212. 1934.

Buley, F. L. and Domingo, C. 1. Effect of water temperature

on rate of Infiltration. Soil Soi. Soc. Amer. Proc.
1943/8:129-131. 1944.
Edlefsen, H. S. Report of the committee on physics of
soil ffloisture. Amer. Geophys. Union Trans 21:414-433.
1940.
Emmons, Howard W. The numerical solution of partial
differential equations. Quar. App. Math. 2*173-195.
1944.
Fireman, M. Permeability measurements in disturbed soil
safflples. Soil Sci. 58:337-354. 1944.
and Bodman, 0. B.
water upon permeability.
1939/4:71-77. 1940.

Effect of saline Irrigation

Soil Sci. Soc. Amer. Proc.

and Magistad, 0. C. Permeability of five

western soils as affected by the percentage of sodium
of the Irrigatioa-water. Araer. Geophys. Onion Trans.
36:91-94. 1945.
Fox, I#. Solution by relaxation methods of plane potential
problems with mixed boundary conditions, (^uar. App,
Math. 2:251-257. 1944.
Free, G. R., Browning, G. M. and Musgrave, G. W. Relative
infiltration and related physical characteristics of
certain soils. U. S. Dept. Agr. Tech. Bui. 729. 1940.

-99-

(27) Frevert, R. Development of a three-dimensional electric

analogu with application to field measurement of
oil permeability below the water table. Ph.D.
thesis, Iowa State College Library. 1948.
(28)< Gardner, fillard The influence of soil characteristics
on drainage and irrigation practices. Soil Sci. Soc.
Amer. Proc. 1936/1:383-392. 1937.
(29)

, Collier, T. R., and Faxr, D. Groundwater,

part I: Fundamental principles governing its physical
control. Utah Agr. Exp. Sta. Bui. 253. 1934.

(30)

Gilles, D. C. The use of interlacing nets for the applica

tion of relaxation methods to problenm involving two
dependent variables. Proc. Roy. Soc, A 193:407-433.
1947.

(31)

Gustafason, Yngve The influence of temperature on the

permeability of soils to water. Lantbrukshogskol
Ann. 8:425-456. 1940.

(32)

Hagen, Q. IJber die Bewegung des Wassers in engen

Oylindrischen Rohren. Ann. Phys. u. Chem. 46:423-442.
1839.

(33)

Hamel, G. Uber Grundwasserstromung.

und Mech. 14:129-157. 1934.

Ztschr. Angew. Math.

(34) Harding, S. 1. and Wood, J. K. Model tests of flow into

drains. Soli Sci. Soc. Amer. Proc. 1941/6:117-119.
1942.
(35)

Harris, A. 1. Effect of replaceable soditim on soil

permeability. Soil Sci. 32:435-446. 1931.

(36)

Hissink, D. J. Kie reclamation of a part of the Zuyder

Zee. Int. Soc. Soil Sci. Trans. 6th Comm. A:177-187.
1932.

(37)

(38)

and Hooghoudt, S. B. Die Bestiaaaung des

Mittleren D-wertes einex Bodenschicht. Int. Soc.
Soil Sci. Trans. 6th Comm. A:34-41. 1932.
Hooghoudt, S. B. Btldragen tot de Kenais van eenige
latuurkundige Grootheden van den Grond. No. 4.
Bepaling van den Doorlaatfactor van den Grond met
Behulp van Porapproeven, e,rs.
. Ond. 42(13). ^. 19 36

(39) Jacob, C. 1. On the flow of water in an elastic artesian

aquifer. Amer. Geophys. Union Trans. 21:574-586. 1940.

-100(40)

King, f. H. Principles and condition of the movements

of groimd water. U. S. Geol. Survey Ann. Rpt.
19:61-394. 1899.

(41)

Kirkham, D. Proposed method for field measurement of

permeability of soil below the water table. Soil
Soi. Soc. Amer. Proc. 1945/10:58-68. 1946.

(42)

Reduction in seepage to soil underdrains

resulting from their partial embedment in, or
proximity to, an impervious substratum. Soil Soi.
Soc. Amer. Proc. 1947/12:54-59. 1948.

(43)

Flow of ponded water into drain tubes in

soil overlying an impervious layer. Amer.
Geophys. Union Trans. 30:In press.

(44)

Unpublished manuscript, Iowa State College,

Ames, Iowa. 1949.

(45)

and van Bavel, C. H. M. Theory of seepage

into auger holes. Soil Sci. Soc. Amer. Proc.
1948/13: In press,

(46)

Kozeny, J. Hydrologische Grundlagen des Dranv^suches.

Trans. Sixth Goram. Int. Soc. Soil Sci. A:42-67. 1932.

(47)

Ueber den Ylasserzu^drang bei Dranfugen.

Wasserkraft uad lasserwirtschaft 28:13-17. 1933.

(48)

Liebmann, H. Die angenaherte Ermittlung harmonischer

Funktionen uad Koaformer Abbildungen, Sitz. Bayer.
Acad. Wiss. Math.-Phys. Klasse, 385-416. 1918.

(49) Lutz, J. F. and Learner, R. 1. Pore-size distribution as

related to the permeability of soils. Soil Sci. Soc.
Amer. Proc. 1939/4:28-31. 1940.
(50)
(51)

Moskovitz, D. The numerical solution of Laplace's and

Poiason*s equations. Quar. App. Math. 2:148-163.

1944.

Muskat, M. The flow of homogeneous fluids through porous

media, McGraw-Hill Book Go., Hew York. 1937.

(52) Kelson, W. R. and Baver, L. D. Movement of water throu^

soils in relation to the nature of the pores. Soil
Soi. Soo. Amer. Ppoc. 1940/5:69-76. 1941.

-101
(53)

Poiseuille, J. L. M. Ejperlmental investigation upon the

flow of liquids in tubes of very small diameter. Roy.
Aoad. Sol. Inst.
Math. Phys. Sol. Mem.
9S432W543. 1846. (frns. by W. H. Herschel in Rheo.
Mem. it1-10. 1940).

(54)

Reeve, R. and Jensen, M. Use of piezometera for groundwater

flow studies and measurement of subsoil permeability*
Agr Eng. In Press.

(55)

Reynolds, 0* Aix experimental investigation of the oircumstaneea which determine whether the motion of water
shall be direot or sinuous and of the law of resistance
in parallel channels. Roy. Soo* (London) Trans.
174:935-983. 1883.

(56)

Richards, L, A. Goneei'ning permeability units for soils.

Soil Sci^ Soc. Amer. Proe. 1940/5s49-53. 1941.

(57)

Discussion of permeability units.

Qeophys* Union Trans. 21*427-428* 1940.

Amer.

(58)

Richardson, Ii. F. The approximate arithmetical solution

by finite differences of physical problems involving
differential equations, with an application to the
stresses in a masonry dam. Roy. Soc. (London) Trans.
A3101307-357. 1910.

(59)

Russell, J. L. Setentifio research in soil drainage.

Igr. Sci. 24:544-573. 1934.

(60)

Seelheim, F. Methoden zur Bestimmung der Durchlassigkelt

des Bodens. Zeit, fur Anal. Ghem. 19i387-418. 1880.

(61)

Shaw, F. S. and Southwell, R. Y. Problems relating to the

percolation of fluids throx^h porous materials.
Roy. Soc. (London) Proc. A178:l-17. 1941.

(62)

Shortley, G., feller, R., and Fried, B. Numerical solution

of Laplace*s and Poisson's equations with application
to photoelasticity and torsion. Ohio State Eng.
Ixp. Sta. Bui. 107. 1940.

(63)

Slichter, G. S, Theoretical investigation of the motion

of ground water. U. S. Geol. Survey Ann. Rpt. 19;
295-384. 1899.

(64)

Smith, R. 1. and Browning, D. R. Influence of evacuation

upon laboratory percolation rates and wetting of
undisturbed soil samples. Soil Sci. 62:243-254.
1946.

Jour.

-102-

Southwell, R. V. Relaxation inethods in theoretical

physics. Oxford Press. 1946.
ferzaghi, Charles Principles of soil meohanioe: III
Deterffiiaation of permeability of clay. Eng.
lews Rec, 95;832836. 19S5.
Tolman, G. F. Groundwater, p.217-219.
lew fork. 1937

McGraw-Hill,

ITan Bavel, C. H. M. and Kirkham D. Field measurement of

soil permeability using auger holes. Soil Sci. Soc.
Amer. Proc. I948sl3. In Press.
f^aksamn, S. A. and Martin, J. P. The role of microorganisms
in the conservation of soil. Science 90:304-305. 1939.
Weir, 1. f. Shape of the water table in tile drained land.
Hilgardi a 31143-152. 1938.
fenzel, I.. I. Methods for determining permeability of
water-bearing materials. U. S. Geol. Survey WaterSupply Paper 887:1-198. 1942.
lyokoff, R, D. and Botset, H. G. The flow of gas-liquid
mixtures through unconsolidated sands. Physics
7:325-345. 1936.

, Botset, I* G., Muskat, M. and Reed, D. ?/.

The measureaient of the permeability of porous media
for homogeneous fluids. Rev. Sci. Inst. 4:394-405.
1933.

-103-

TI. AQUOfldDSSlffilT

Til aatlior wislies t express htm sincere appreciatioa

to those iho aided diArli^ the oomdmct of the experiment.

The

ontrllwtioas of the followlog are eepecially recognized.

ttr, Don Kirliliaa, in oharge of aajor work, for his
gmidaace aM assistano.
Dr. J. J. L. Hinris&sea, Dr, t, T. Sarle, Dr. 1. H.
Allaway, Dr. Q. M. Broming, and Dr. R. K. FreTert, members
of the oofflffiittee# for -ttielr interest and suggestions.
Dr. R. E. Gaskell for his aid in the derivation of
ep.ations for aafflerieal analysis and for his interest and
encourageffient.